(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Value of mussel beds to sport fisheries, Project F-80-R : final report"

<//31 rs 



/T 



ILLINOIS 
NATURAL 
HISTORY 
SURVEY 



=\ 



Value of Mussel Beds to Sport Fisheries 

Final Report, F-80-R 



Center for Aquatic Ecology 



Philip B. Moy and Richard E. Sparks 



December 1 991 



k 



Aquatic Ecology Technical Report 91/17 




DISTRIBUTED TO: 



copies sent to Larry Dunham (IDOC) 

copy sent to Mike Sweet (IDOC) 

copy sent to Bill Bertrand (IDOC) 

copy sent to Aquatic Office CU 

copy sent to Doug Blodgett (LTRM-Havana) 

copy sent to Chuck Theiling (LTRM-26) 

copy sent to Phil Moy 

copy sent to INHS Library CU 

copy sent to Dr. Larry Jahn (WIU) 

copy sent to Dr. Richard Anderson (WIU) 

copy in Forbes Biological Station Library 

copy sent to Pam Thiel (EMTC) 

copy sent to Ken Lubinski (EMTC) 

copy sent to Kevin Cummings (INHS-CU) 



Value of Mussel Beds to Sport Fisheries 

Project F-80-R 

Final Report 



Philip B. Moy and Richard E. Sparks 

Illinois Natural History Survey 

Forbes Biological Station 

P.O. Box 599 

Havana, Illinois 62644 



December 1991 



Digitized by the Internet Archive 

in 2010 with funding from 

CARLI: Consortium of Academic and Research Libraries in Illinois 



http://www.archive.org/details/valueofmusselbedOOmoyp 



Value of Mussel Beds to Sport Fisheries 
Project F-80-R 
Final Report 



Philip B. Moy and Richard E. Sparks 

Illinois Natural History Survey 
Forbes Biological Station 
P.O. Box 599 
Havana, IL 62644 



December 1991 




^^2 



Dr. Philip B. Moy 
Principal Investigator 



(^~*\ZK~ \ 




Dr. David P. Philipp, 
Center Head 



Dr. Richard E. Sparks / 
Co-Principal Investigator 



DISCLAIMER 

The findings, conclusions, and views expressed herein are those of 
the researchers and should not be considered as the official position of 
the United States Fish and Wildlife Service or the Illinois Department 
of Conservation. Mention of product names does not imply endorsement by 
the sponsoring agencies or the researchers. 



ABSTRACT 

The larvae, or glochidia, of freshwater mussels are obligate 
parasites on fish, including sport fish. The host fish provides a 
means of dispersal and nutrients that enable the glochidia to transform 
into juvenile mussels. Although some mussels use visual lures (mantle 
flaps or glochidial packets that look like minnows, worms, or other 
types of forage) to attract fish, many mussels that are common in the 
Midwest do not have such obvious means of attracting fish. The purpose 
of this research was to determine whether one forage fish (fathead 
minnow) and five common sport fishes (channel catfish, largemouth bass, 
bluegill, white crappie, and sauger) were more attracted to mussel beds 
than to other types of substrate, and if they were, to determine the 
basis for that attraction. 

Individual sport fish and schools of fathead minnows were given a 
choice of two substrates in laboratory tanks. Substrates included 
mussels and cobbles with naturally occurring attached algae and 
invertebrates; mussels with cleaned shells; clean cobbles; and bare 
sand. The mussels were partially buried in sand in a natural position, 
with their pseudosiphons and the posterior portions of their shells 
above the sand surface. The cobbles were approximately the same size as 
the mussels and similarly placed in sand. Each fish was exposed to the 
same substrate (bare sand or clean cobbles) on both left and right sides 
of the tanks during a control trial to quantify its side preference. 
Software running on Apple IIGS computers linked to image digitizing 
cameras determined the location of the fish every 2 seconds during the 
day and every 4-8 seconds at night, when the room lights were 



automatically dimmed and longer exposure intervals were required by the 
cameras. A substrate was defined as preferred only if the fish was 
located more frequently over the side with that substrate than during 
the control trial . 

Fish were located more frequently over mussels and cobbles with 
attached algae and invertebrates than over clean cobbles or sand. The 
invertebrates were an obvious attraction that the fish consumed within a 
few minutes. The preference for mussels and cobbles usually diminished 
during the 1- to 3-day duration of the trials, a behavior that might be 
attributed to an initial food reward, followed by a gradual decline in 
unrewarded searching of the same substrate. However, there was some 
other attraction associated with the mussels themselves because there 
was no statistically significant difference (P > 0.05) in preference 
between mussels with cleaned shells and mussels or cobbles with attached 
algae and invertebrates. When handled, some of the mussels aborted 
glochidial packets that the fish readily consumed. The aborted packets 
contained 55-77% protein on a dry weight basis and the embryonic larvae 
had scarcely any shell material. The protein content dropped to 20% in 
mature packets, because shell material comprised a greater portion of 
the total mass. Both immature and mature packets appear to be a 
nutritious food source that presumably can be digested by the fish--the 
only glochidia that survive are the ones that attach externally to the 
gills or fins of the fish. In contrast to the glochidial packets, the 
feces and pseudofeces produced by the mussels did not appear to be very 
nutritious for the fathead minnow, a species known to consume detritus; 
fathead minnows maintained in tanks with mussels lost as much, or more 



ill 



weight than fish held alone. In predator-prey trials, the minnows were 
twelve times more vulnerable to sauger when the minnows were over bare 
substrate than when over cobbles, and consequently they spent twice as 
much time over the cobbles. The minnows presumably would use mussels 
similarly as a refuge from predation, although time did not permit 
testing that hypothesis in additional predation trials. The possibility 
that mussels release odors that attracted their fish hosts was not 
investigated, but should be the subject of additional research. During 
these preference trials, none of the mussels displayed lures that could 
have visually attracted the fish. 

These results suggest that mussels serve as both direct and 
indirect sources of forage for game fish. The glochidial packets 
released by the mussels may provide a seasonally abundant food reward 
for the fishes that disperse the larval mussels. Young game fish may 
consume invertebrates that colonize mussel shells, or the link may be 
from invertebrates to small fish, such as fathead minnows, to the 
piscivores. Small fish also may concentrate in mussel beds to avoid 
predators or currents. Mussels can serve as solid substrate and refuges 
for other invertebrates and forage fish because most native mussels 
continually expose a portion of their shells above the sediment surface, 
rather than completely burying themselves. Although cobbles can perform 
the same functions, cobbles can be covered by silt or sand whereas the 
living mussel actively maintains its position at the sediment surface. 
The solid structure provided by mussel beds is likely to be most 
critical as a substrate in alluvial rivers otherwise dominated by 
shifting deposits of sand or mud. 



ACKNOWLEDGEMENTS 

Financial support for this project was provided by U.S. Fish and 
Wildlife Service Sport Fish Restoration funds administered by the 
Illinois Department of Conservation. 

Administrative support for this project was provided by the 
following individuals: Mr. Robert Adair of the U.S. Fish and Wildlife 
Service, Mr. Larry Dunham, Mr. William Bertrand, Mr. E. Butch Atwood, 
and Mr. Michael Sweet, of the Illinois Department of Conservation, and 
Dr. Lorin Nevling, Dr. David Philipp, and Dr. Stephen Havera of the 
Illinois Natural History Survey. 

Mr. Robert F. II lyes wrote the FishWatcher programs we used to 
record, summarize, and analyze the position of fish in the preference 
tanks. He also recommended, assembled, and tested the video and 
computer hardware. Dr. Jerry Colliver and Dr. Steven Verhulst of 
Southern Illinois University School of Medicine developed the 
statistical model for analysis of the substrate preference data. 

Many individuals associated with various agencies have aided this 
project by providing materials or labor: Mr. Scott Stuewe, Mr. Steven 
Krueger, Mr. Alan Brandenburg, Mr. Dan Sallee, Mr. Ed Walsh, Mr. Robert 
Schanzle, Mr. Rod Horner, Mr. Larry Durham and Mr. Rudy Steinhauer of 
the Illinois Department of Conservation; Mr. Doug Blodgett, Ms. Cammy 
Smith, Mr. Eric Hopp's, Mr. Paul Raibley, Mr. Frank Dillon, Mr. Brian 
Todd, Dr. Lewis Osborne, Ms. Liesl Mensinger, Ms. Katie Roat and Mr. 



Larry Gross of the Illinois Natural History Survey; and Dr. Roy 
Heidinger, Mr. Bruce Tetzlaff, and Dr. David Bergerhouse of the 
Cooperative Fisheries Research Laboratory, Southern Illinois University 
at Carbondale. Mr. Frank Budyn generously provided the video camera we 
used to record the yellow sandshell mantle flap display. 



vi 



TABLE OF CONTENTS 

DISCLAIMER i 

ABSTRACT i i 

ACKNOWLEDGEMENTS v 

LIST OF TABLES ix 

LIST OF FIGURES x 

LIST OF APPENDICES xi i i 

INTRODUCTION. Job 101 . 1 1 

Need for research 1 

Parasitism of fish 3 

Mussels as forage 9 

Mantl e f 1 ap 1 ures 10 

Glochidial congluti nates 10 

Other i nvertebrates and mussel s 11 

Feces as forage 13 

Pseudofeces 14 

Pseudof eces as forage 14 

Sedimentation of organic matter 16 

Habitat permanence 17 

Interstitial habitat 17 

Substrate heterogeneity 19 

Bioturbation 19 

Summary of literature 21 

RESEARCH QUESTIONS 23 

APPROACH 24 

PREFERENCE STUDY. Jobs 101.2 - 101.6 27 

Methods and materi al s 27 

Test chambers 27 

Water Qual ity 29 

Fish species 29 

Acclimation 30 

Control Tri al s 30 

Variable substrates 30 

Data collection 32 

Data analysis 35 

Substrate preference resul ts 37 

Colonized mussels 37 

Scrubbed mussel s 40 

CI ean rocks 40 

Colonized rocks 45 

Dead shells 45 

Substrate preference discussion 48 



vn 



TABLE OF CONTENTS cont'. 



PREDATOR-PREY TRIALS. Job 101.6 53 

Methods 53 

Analysis 53 

Predator-prey results and discussion 56 

FEEDING TRIALS. Job 101.6 61 

Materials and methods 61 

Water del ivery 61 

Water treatment 63 

Water qual ity 65 

Treatments 65 

Species 65 

Repl icates 66 

End of trial 68 

Sediment protein 69 

Statistical analysis 70 

Results of feeding trials 71 

Fish weight change 71 

Sediment protein 73 

Sediment weight 87 

Water qual ity 87 

Mussel weights 93 

Feeding trials discussion 97 

SUMMARY AND CONCLUSIONS 104 

Laboratory studies 105 

Importance of mussel beds to fishes 109 

Recommendations. . .-.■ Ill 

LITERATURE CITED 116 

APPENDICES 134 



vm 



LIST OF TABLES 
Table Page 

1. Fish which associate with mussel beds 5 

2. Invertebrates found in association with mussels 12 

3. Variable and control substrates used with each species... 31 

4. Mussel species used as substrate in the preference trials.. 33 

5. Summary of (day, night) results from the substrate preference 

trials for all species 49 

6. Summation of responses to the substrates for all species (from 

tabl e 5) 50 

7. Mean time (T) spent over rocks and sand by sauger and fathead 

minnows, percent of minnow capture and capture per fish-minute 
by substrate in the predator-prey study 57 

8. Conditions and treatments for feeding trials 68 

9. Correlation and regression coefficients of biological and 

environmental variables with percent change in fish weight.. 94 

10. Percent protein content of glochidial conglutinates as 
determi ned by two methods 141 

11. Number of mortalities and survivors of mussels frozen in 

outdoor tanks 146 

12. Mean physico-chemical parameters measured at a non-mussel area, 

a mussel sanctuary and an exploited mussel bed in pool 26 of 
the Mississippi River 150 

13. Species composition at each site in pool 26 of the Mississippi 

River 153 

14. Regressions (r) between total daily number of fish captured 

over the four day period and physico-chemical parameters at 
the three sites 159 



IX 



LIST OF FIGURES 
Figure Page 

1. Top view of experimental tank and associated apparatus 28 

2. End-on view of substrate preference tank 34 

3. Fish response to colonized mussel substrate on the first day of 

the preference trial 38 

4. Fish response to colonized mussels after three days 39 

5. Fish response to scrubbed mussels on day one 41 

6. Fish response to scrubbed mussels after three days 42 

7. Fish response to clean rocks on day one 43 

8. Fish response to clean rocks after three days 44 

9. Response of fish to colonized versus clean rock 46 

10. One-day and three-day response of fish to dead shells 47 

11. Frequency of substrate preference for all fish species in the 

substrate preference trials 51 

12. Comparison of percent time spent over rock and sand by sauger 

and fathead minnow in the predator-prey study and the 
substrate preference tri al s 59 

13. System design for the mussel feces feeding trials with fathead 

minnows 62 

14. Detail of one of the thirty, five-gallon tanks used for the 

feeding trials 63 

15. Percent weight change of fish in the feeding trials 72 

16. Mean sediment protein concentration (ppt protein/sediment) for 

each treatment in the feeding trials 74 

17. Regression of percent fish weight change with sediment protein 

concentration (ppt protein/sediment) 76 

18. Regression of percent weight change of fish held alone with 

sediment protein concentration (ppt protein/sediment) 77 

19. Regression of percent weight change of fish held with thick- 

shell mussels and sediment protein concentration (ppt 
protein/sediment) 78 



20. Regression of percent weight change of fish held with thin- 

shell mussels with sediment protein concentration (ppt 
protein/sediment) 79 

21. Regression of sediment protein concentration (ppt protein/ 
sediment) with mean daily water exchange (gallons/day) 80 

22. Net sediment protein concentration (ppt protein/sediment) for 

each treatment in the feeding trials 82 

23. Regression of net sediment protein concentration (ppt protein/ 

sediment) and mean daily water exchange (gallons/day) 83 

24. Regression of percent fish weight change and net sediment 

protein concentration (ppt protein/sediment) 84 

25. Regression of percent fish weight change and mean daily water 

exchange (gallons/day) 85 

26. Net sediment protein concentration (ppt protein/sediment) per 

pound dry tissue weight for each treatment in the feeding 
trials 86 

27. Mean dry sediment weight collected from each treatment in the 

feeding trials 88 

28. Concentration (ppm) and percent saturation of dissolved oxygen 

for each treatment in the feeding trials 89 

29. Mean turbidity in formazin turbidity units (FTU) for each 

treatment in the feeding trials 90 

30. Mean un-ionized ammonia concentrations for each treatment in 

the feeding trials 92 

31. Percent fish weight change in the feeding trials compared with 

mussel to fish weight ratio, temperature and daily water 
exchange (gal 1 ons/day ) 98 

32. Repositioning movement of a yellow sandshell mussel to assume 

the headstand position for the mantle flap display 136 

33. Mean number of each species of fish captured in pool 26 of the 

Mississippi in non-mussel, mussel sanctuary and exploited 
mussel bed areas 154 

34. Total number of each species of fish captured in pool 26 of 

the Mississippi in non-mussel, mussel sanctuary and exploited 
mussel bed areas 156 



XT 



LIST OF FIGURES cont'. 
Figure Page 



35. Mean length of each species of fish captured in pool 26 of the 

Mississippi in non-mussel, mussel sanctuary and exploited 
mussel bed areas 157 

36. Combined length - frequency distributions for all species at 

the non-mussel, mussel sanctuary and exploited mussel bed 
sites in pool 26 of the Mississippi River 158 



xn 



LIST OF APPENDICES 
Appendix Page 

A. Yellow sandshell mantle flap display 134 

B. Protein content of glochidial conglutinates 140 

C. Filtration rates of some marine and freshwater bivalves 144 

D. Survival of frozen mussels 145 

E. Total time (fish-minutes) spent over sand and rock in the 
predator-prey trials by sauger and fathead minnows 147 

F. Calculation of percent dry tissue weight from whole wet body 

weight for mussels used in the feeding trials 148 

G. Methods and results of field sampling at pool 26 of the 

Mississippi River 149 



xm 



INTRODUCTION 

Today habitat degradation in large Midwestern rivers threatens 
the health of populations of many aquatic organisms. Siltation 
specifically threatens to cover cobble and gravel substrates which 
form spawning and feeding areas for riverine fishes. These areas 
cannot be maintained free of silt without additional human 
intervention. 

Like cobble, mussels provide a firm substrate for attachment of 
invertebrates (Anderson and Vinikour 1984) and spawning of fishes 
(Pitlo 1989). Unlike cobble, mussels maintain their position despite 
sedimentation or scouring and produce feces and pseudofeces that are 
consumed by other invertebrates (Izvekova and Lvova-Katchanova 1972). 

Need for research. Because mussels depend upon fish for 
completion of their life cycle (Ellis 1929) we know that fish 
associate with mussels at least during the period of glochidial 
release. It is not known what effect, if any, elimination of mussel 
beds through loss of suitable habitat, deteriorating water quality, 
and overharvest (Fuller 1978) may have on fish. 

Mussels have been commercially harvested for various reasons 
over the last century. In the latter part of the nineteenth century 
mussels were harvested for food and for the pearls infrequently found 
in their mantle cavity. Later, in the 1890's and early part of this 
century mussels were harvested for the pearl button industry. Today 

1 



mussels are harvested for their shells which are ground into spheres 
and implanted in oysters as nuclei for cultured pearls (Fuller 1974). 

The vast natural mussel beds in Midwestern United States rivers 
supported a thriving pearl button industry for many years, though 
mussel harvests were already beginning to decline in the early 1900's 
(Eckblad 1986). The decline in populations and the economic importance 
of mussels fostered the need for research on life history and 
propagation with the intent to restock mussel beds to support the 
button industry (Fuller 1974, Coker et al . 1921). The result was a 
large body of information regarding the life habits of these 
animals. Subsequent to World War II the advent of plastics resulted in 
the closing of the pearl button factories and elimination of the 
industry. Unfortunately this also resulted in elimination of the 
primary reason for mussel research (Starrett 1971, Fuller 1974). 

There have been few investigations of fish substrate preference 
that have produced conclusive results. Some authors suggest that fish 
substrate preference is based upon forage organisms associated with 
the substrate (Rankin 1986), though interstitial space may be 
important for small fish (Sechnick et al . 1986, DeMarch 1976). Work 
performed on substrate preference of invertebrates has produced mixed 
results, but again authors suggest food is influential in substrate 
selection (Egglishaw 1964), though particle size and interstitial 
space (Cummins and Lauff 1969), and current velocity and dissolved 
oxygen are suggested as factors as well (Eriksen 1966). 



The majority of information on fish-mussel interactions comes 
from the mussel literature pertaining to parasitism of fish by mussels 
(Howard and Anson 1922, Coker et al . 1921), identification of host fish 
for endangered mussels (Holland-Bartels 1990, Miller et al . 1986, Zale 
and Neves 1982, Sephton et al . 1980, Stern 1978), use of fish to 
propagate mussels (Coker et al . 1921, Howard 1914, 1917), and species 
of fish caught over mussel beds (Wilson and Clark 1912). Other fish - 
mollusk information pertains to fish consumption of snails or mussels 
as forage (Bennett and Gibbons 1972, Forbes and Richardson 1908). 
Some of the early papers speculate on reasons for the presence of fish 
over mussel beds and cite food or forage as the probable cause. Yet no 
one has closely examined this relationship to clearly identify whether 
fish are more attracted to mussels than to non-living substrates. 

Parasitism of fish. Larval mussels were originally thought to be 
parasites infesting the gills of mussels, and were given the name 
( Glochidium parasiticum ) . In 1832 Carus proved that these parasites 
were actually larvae of the mussel itself, though the term glochidium 
is still used in reference to the larval stage (LeFevre and Curtis 
1912). The dependence of mussels on fish for completion of their life 
cycle was discovered in 1866 when Leydig identified parasites on the 
gills of fish as glochidia (Ellis 1929). The glochidia of nearly all 
species of unionid mussels must parasitize a fish for a short time 
after leaving the female mussel (Clark and Stein 1921) and often 
display a high degree of host specificity for successful completion of 
larval development (Howard 1914, 1917, Howard and Anson 1922, Zale and 
Neves 1982). The larval mussel attaches to the gill lamellae or fins 



of the fish and remains there for a period of days or weeks depending 
on temperature (Davenport and Warmuth 1965, Zale and Neves 1982), 
mussel species and developmental stage of the glochidium (Howard and 
Anson 1922). During this period of parasitism the fish provides the 
young mussel with nutrition for metamorphosis (Arey 1932, Ellis 1929) 
and with a means of dispersal (Coker et al . 1921, Starrett 1971, 
Fuller 1974). Encysted larvae metamorphose to the juvenile stage at 
various rates and apparently do not drop off en masse; glochidia 
encysted on one fish from one mussel may excyst over a period of two 
weeks or more (LeFevre and Curtis 1912, Howard 1922). 

Many fish are hosts for freshwater mussels. In their 1912 survey 
of the mussel fauna of the Kankakee River basin Wilson and Clark made 
seine hauls over mussel beds and recorded the various species of fish 
they captured. Similarly, Wiles (1975) used seine hauls, dip nets and 
electrofishing to sample the fish over the mussel beds he studied in 
Nova Scotia. In their study of lampsilid fish hosts, Zale and Neves 
(1982) captured several species of fish over mussel beds in Big 
Moccasin Creek, Virginia. Table 1 provides a compilation of fish 
captured in the above studies along with fish known to be glochidial 
hosts . 



Table 1. Fish which associate with mussel beds. * Denotes fish 
not known to be glochidial hosts which have been captured over mussel 
beds. Unmarked fish are known mussel hosts. Numbers correspond to 
forage items, upper and lower case letters refer to citations for 
association with mussels and diet respectively. See end of table for 
key. 



Species 



Diet 



Petromyzontidae 

Petromvzon marinus A 
Acipenseridae 

Scaphirhvnchus platorhvnchus A 

Lepisosteidae 

Lepisosteus p1atostomus A 
L. spatula A 
Lj. osseus A 

Amiidae 

Ami a cal_va A 
Anguill idae 

Anquilla rostrata fl 

Clupeidae 

Alosa chrvsochlori Sp 
/L pseudoharenqus u 'jl 
Dorosoma cepedianunr 

Salmonidae 

Oncorhynchus mvkiss A 

0. nerka A 

(L trutta A 

(L tschawy_tscha A 

Salve! inus fojvtjnalis A 



8 (parasitic) 3 ' b 



3-5< 



8 a,c 
8 a,c 
8 a,b 



8, crayfish a ' b 



3-8, crayfish, snails 1 



3-8 c 



1,2, amphipod: 



2,9 



2r8, snails, crustaceans 3 '" 

2 h 

3-8, crustaceans 3 ' , 

3-7, crustaceans (in fresh water) 



3-8, crustaceans 



a,b 



Table 1 cont' 



Esocidae 



Esox americanus * e 
L. lucius B,L 

Cyprinidae 

Campostoma anomalu m *^ 

Carassius auratus ^ 

Chrosomos eos *^ 

Cyprinus carpio A 

Ericymba buccata * B 

Hybopsis amblops * B, £ 

Nocomis biquttatus * B 

\L_ micropoqon * 

Notemigonus crysoleucas A » B » ^ 

Notropis ardeg s A 

iL blennius * B 

N_s. coccoqenis *^ 

N^ cornutus * B >C,D 

PL. qalacturus * c 



N. heterodon * 



hL heterolepis * u 
iL leuciodus * c 
IL rubellus * c 
tL spilopterus * B 
fL. teloscopus *~ 
fL umbratilis * B 
N. whipplei * B 
Phenacobius mirabili 



£ r 

Pimephales no tatus * D ' b 

P_t promelas ^" 
J\ viqilax * B 
Rhinichthys atratulus * c 
R. osculuT * 

Richardsonius eqreqius A 
Semotilus atromaculatus B 

Catostomidae 

Catostomus commersoni ' C '^'^ 
Ll tahoensir * ~ 
Carpiodes vel i f er * 
Erimvzon sucetta 6 
Hypentelium nigricans 6 '^ 6 
Moxostoma macro! epidotutir 
[L duquesnei * B »^ 



6,8, amphipods, isopods c 
6,8 C 



5 ' 9 , 

1-7 h 

l-7 b 

1-7, 9 a 

2,3,5.9 C 

6,7*> c 

1,4-6 C 

1,2-5,8, snails, crayfish b 

1,2,6,7,9, molluscs 

1, "animal material" 3 

1,2,6 C 

l-5,7-9 d 

1,2,5 C 
1-7 3 

1,3, 5,7, 9 d 
1-5, 7 d 

Unknown 3 

3-5,9 c ' d 
l-6,9 c ' d . 
l,5,7,9 c ' d 
1, 2-5.7, 9 d 
1,4.5 6 

l-5,8,9 c ' d 



3-5, molluscs, crustaceans 

2,3-5,9, Sphaerium . Lymnea c 

"benthic invertebrates" 3 
3,5,9, mussels 
2-5,9, Sphaerium , Snails 
6,9, molluscs 3 



B,E 



Table 1 cont' . 

Ictaluridae 

Ictalurus melas ^ 

L. natal is C 
Jj. nebulosus A 
Xi punctatus A 
Noturus gyrinus 
Pvlodictus olivarus " 

Cyprinodontidae 

r n 
Fundulus di aphanus D,u 

L. dispar *° 

£,. zebrinus A 

Poecil iidae 

G ambus i a affinis 



<,A 



Atherinidae 

Labidesthes sicculus * B 

Gasterosteidae 

Apeltes quadracus 
Culea inconstans A 
Gasterosteus aculeatus u 
Punqitius punqitius ^ 

Percichthyidae 

Morone americana 
M^ chrysops A 

Centrarchidae 

r r 
Ambloplites rupestris p ' 

Lepomis cyap e llus 

L, gibbosus A ' D 

L. gul^suift' ' ' 

L. humi1is A 

L. macrochiru s A ' B 

U megalotis" 

Micropterus dolomieui ' ' ' 

Mi salmoides '^ 11 '" 

Pomoxis annularis ^'* 1 

Pi niqromacu1atus A 



1,3-8, mussels, snails, 

crustaceans 
1-8 C 

1-8, mussels, snails 
1-8, snails, Anodonta 
2-5, amphipods. isopods 
8, crustaceans 3 ' 



2,6,7, snails, amphipods, odonates 
1,2,6,7, snails 3 ' 



2-6 c 



2,5,7 ( 



1,2 D 

1-5, snails 3 ' 

3 ' 8 k 
2-6 a ' b 



2-5,8 



aTb 



2-8 



2-8 a crustaceans 3 ' 

2-8 b 

2-8. molluscs 3 ' ' 

2-8 3 ' 

2-8 a 

1-8- molluscs"' 

2-8 b 

2,3,6-8, crustaceans 3 ' 

2,3,6-8, crayfish 3 ' 

3-6..8, crustaceans 3 ' ' 

2-8 & 



Table 1 cont' 



*B 



Percidae 

r 

Etheostoma bl ennoides *° 

£j_ caeruleum *" 

L. exile b 

L. flabellare 1 

L microperca 

Ei nigrum " 

Ei rufilineatunr 

L simoterunT **' 

Perca flayescens A 'p' IJ ' t ' 1 

Percina caprodes *^ 

Pi phoxocephal a * B 

Pi maculata *°~ 

Stizostedion canadense A 

S^. vitreum vitreunT 

Sciaenidae 

Aplodinotus qrunniens A 

Cottidae 



Cottus bairdJ A ' B 
C. carol inae* 1 



»?c 



2-5. benthos a ' b 

3-5 3 

2-5, amphipods, snails 3 ' b 

3,4,6, coleoptera larvae b 

2.3.5. amphipods 3 ' b ' c 
2,3,5 b 

1,2, drift* 

benthos 

3-5,8 b '° 

2.5.6, molluscs 3 ' 
3,5,6, odonates 3 ' 
3,5,6, odonates 3 'F 
2-6,8, crayfish 3 ' 13 
2-8, crayfish 3 ' 



2-6,8, mussels, snails, 
crustaceans c 



3-5, crayfish 3 ' 
6,8, crustaceans 3 ' 



Key to forage items: 1 - algae; 2 - zoopl ankton ; 3 - 

Ephemeroptera nymphs; 4 - Trichoptera larvae; 5 - Diptera larvae; 6 - 

adult aquatic insects; 7 - adult terrestrial insects; 8 - fish; 9 - 
bottom ooze (diatoms). 

Key to fish species citations: A - Fuller (1974); B - Wilson and 
Clark (1912); C - Zale and Neves (1982); D - Wiles (1975); E - Coker 
et al . (1921); F - Davenport and Warmuth (1965); G- Howard and Anson 
(1922); H - Arey (1932); I - Tedla and Fernando (1969). 

Key to diet citations: a - Smith (1979); b - Scott and Crossman 
(1973); c - Forbes and Richardson (1908); d - Starrett (1950); e - 
Segler (1963); f - Zale and Neves (1982). 



Mussels as forage. Most of the fish species , 1 isted in Table 1 
that associate with mussel beds do not consume mussels as forage and 
many of these species are not known to be glochidial hosts. Minnows 
and darters present over mussel beds feed on invertebrates, vegetation 
and "bottom ooze" (diatoms). Some predacious drift feeders will feed 
on glochidia but there is no indication of selection for glochidia 
(Zale and Neves 1982) . Zooplankton comprise the initial food of larval 
sport fish. As the young fish grow, insect nymphs, larvae and small 
fish comprise an increasing portion of their diet (Gerking 1962, Ney 
1978) and as larger adults, sport fish will consume minnows, darters 
and young of their own and other fish species (Forbes and Richardson 
1908, Smith 1979, Scott and Crossman 1973). 

A few fish species, notably the freshwater drum ( Aplodinotus 
grunniens ) , channel catfish ( Ictalurus punctatus ), blue catfish (L. 
furcatus ) and the redear ( Lepomi s microlophus ) regularly consume 
mussels as forage (Forbes and Richardson 1908, 1920, Howard 1913, 
Wilbur 1969); other species such as the largemouth bass ( Micropterus 
salmoides ) may occasionally consume mussels (Bennett and Gibbons 
1972). Yet most fish known to be hosts for mussels are not 
particularly noted for their consumption of mussels. Fuller (1974) 
suggests that a mutualistic relationship exists between fish and 
mussels because fish which have been infected with glochidia gain 
resistance to parasitic copepods. Howard (1913) suggests a 
relationship similar to that between plants and pollinating insects 
may exist for fish and mussels and that mussel beds are attractive to 
fishes because of the associated invertebrate life in the vicinity. 



Coker et al . (1921) also speculated that food may be the "clue to 
unraveling the mystery" as to why fish are found near mussel beds. 

Mantle flap lures. The existence of lures used by mussels to 
attract potential host fish further suggests that glochidial hosts are 
foraging in areas inhabited by mussels. Several species of Lampsilis 
possess mantle flaps which mimic the form and movements of a small 
fish (Kraemer 1970, Harman 1970, Wickler 1978, Welsh 1969) and may 
serve to attract their piscivorous, sight-feeding hosts. 
Lampsil isteres displays the marsupium (the gills containing the 
glochidia) and the mantle flaps when the glochidia are ready to 
parasitize a host fish (Appendix A). Similarly, during the period of 
glochidial release, Villosa nubulosa emerges completely out of the 
substrate, gapes the valves and fully extends its foot out of the 
shell (Zale and Neves 1982). Smallmouth bass (M^ dolomieui ) , 
largemouth bass, bluegill ( Lepomi s macrochi rus ) . rock bass 
( Ambl oolites rupestris ), white crappie ( Pomoxis annularis ), yellow 
perch (Perca flavescens h walleye ( Stizostedion vitreum vitreum ) and 
sauger (S^. canadense ) are among the known hosts for these mussels 
(Clarke 1981, Mathiak 1979, Fuller 1974, Waller et al . 1985)--all are 
sight-feeding predators. 

Glochidial conqlutinates. An additional example of a lure to 
attract foraging fish is found in the Arkansas fanshell ( Cyproqenia 
aberti ) which forms red, worm-like glochidial conglutinates which 
protrude from the shell opening (Chamberlain 1934). The conglutinates 
are composed of thousands of glochidia held together in a mucous matrix 

10 



with mature glochidia on the outside and immature glochidia on the 
inside. Chamberlain observed these worm-like projections were readily 
consumed by fish. When the fish picks up the packet, mature glochidia 
on the surface of the packet break free in the buccal cavity of the 
host, pass into the gill chamber and attach to gill filaments. The 
remainder of the packet is consumed by the fish. Chamberlain 
speculated that the congl utinates resembled tubificids or other 
bottom-living worms; chironomid larvae are known to associate with 
mussel shells (Beedham 1970, pers. obs.) and would likely have a 
similar appearance. Often when mussels are disturbed, as occurs with 
collecting, they expel glochidial conglutinates. These packets are 
subcyl indrical or flattened in cross section with pointed or blunt 
ends and are readily consumed by bluegill (pers. obs.). 

Other invertebrates and mussels. Non-bivalve macroinvertebrates, 

as indicated above, are food for many fishes and are also found in 

association with mussels. These invertebrates find food and firm 

substrate for attachment in mussel beds. Trichopteran and chironomid 

larvae have been found on mussel shells (Anderson and Vinikour 1984 

i 
and Beedham 1970) and Driscoll and Brandon (1973) in their study of 

fossil sediments found a greater abundance and diversity of suspension 

feeding invertebrates on bottom sediments which have higher 

concentrations of dead shells; material which may serve as attachment 

sites. Types of invertebrates which have been found in association 

with mussels appear in Table 2. 



11 



Table 2. Invertebrates found in association with mussels. Lower 
case letters refer to references as follows: a - Coker et al . (1921), 
b - Sephton et al . (1980), c - Anderson and Vinikour (1984) and d - 
Beedham (1970). 



Turbellaria 3 

Bryozoa a 

Mollusca 
Gastropoda 
Viviparidae 
Vivipara a 
Pleuroceridae 
Pleurocera 3 
Pelecypoda 
Sphaeriidae 
Musculium a 

Annelida 
Oligochaeta . 
Enchvtraeus 
Hirudinea 
Placobdella a 



Crustacea 
Decapoda 
Cambarus a 



Insecta 
Plecoptera a 

Ephemeroptera 
Heptagenia a 

Odonata 
Gomphus 3 
Arqia d 
Neurocordulia a 

Trichoptera 
Hydropshychidae 
Hvdropsvche a 
Leptoceridae 
0ecetis c 

Coleoptera 3 

Diptera 
Chironomidae 

Chironominae. 
Harnischia f^ 
Tanvtarsus " 
Micropsectra " 
Dicrotendipes * 
Polvpedilunr 

Tanypodinae 
Procladius " 

Orthocladiinae^ 
Metriocnemus d 



12 



Sephton et al . (1980) found a greater numerical abundance of 
invertebrate organisms in areas inhabited by mussels than in areas 
devoid of mussels in a New Brunswick reservoir and theorized the 
increase was in response to an increased food source. Sephton and his 
coworkers identified a positive association between Procladius , an 
omnivorous chironomid, and mussels. These workers suggest the positive 
association could be an indirect response of a predator ( Procladius ) 
to higher prey densities in the form of detritus feeders in the 
vicinity of the bivalves. 

Like cobble, there is an organic coating on mussel shells, 
composed of bacteria, attached algae and bryozoans, but mussels also 
produce feces and pseudofeces which can provide forage for aquatic 
insect nymphs and larvae occupying the "scraper" or "collector" 
functional groups outlined by Merritt and Cummins (1984) and Wetzel 
(1983). These groups include Insecta orders Ephemeroptera, 
Trichoptera, and Diptera which would be found in lotic habitats. The 
detritivorous "collectors" in particular could feed upon mussel feces 
and pseudofeces and the protein in these materials may enhance the 
forage value of sediments and detritus in the region of the mussel 
bed. 

Feces as forage. Detritivorous invertebrates consume their own 
feces as well as that of other species (Hynes 1970). The isopod 
Asel 1 us aquaticus benefits from the consumption of grass carp 
( Ctenopharygodon idella ) feces (Petridis 1990). Grass carp consume 
macrophytes and break down the material into smaller particles the 



13 



isopod can more easily digest. It is reasonable to suppose that this 
isopod could benefit, from the consumption of mussel pseudofeces as 
well since fecal matter produced by filter feeders may be further 
processed by "collector" organisms in the stream (McCul lough et al . 
1979). Bacteria are relatively sparse on detritus until it has passed 
through the digestive tract of an invertebrate (Hargrave 1976). Taylor 
and Roff (1984) found food quality of detritus was five times greater 
in downstream reaches than in the head waters due to bacterial 
colonization of the detritus. 

Pseudofeces. Mussels are not selective in their ingestion of 
particles; all particles small enough to be drawn into the incurrent 
siphon are agglutinated into mucous strings. Mussels consume decaying 
organic matter and animal plankters (Fuller 1974, Burky 1983) but may 
also consume phytopl ankton (Carver and Mallet 1990, Shpigel and 
Fridman 1990). As they feed, mussels filter several liters of water 
daily and in doing so sediment planktonic matter (Leff et al . 1990). 
When the concentration of food material exceeds that required for 
maintenance or growth, some of the food is egested, prior to 
digestion, as pseudofeces (Winter 1978). 

Pseudofeces as forage. Izvekova and Lvova-Katchanova (1972) 
demonstrated that the feces and pseudofeces produced by Dreissena 
polvmorpha provide a more nutritious forage for chironomids than the 
same material which has not been agglutinated by mussels and suggest 
that the mucopolysacharride coating on the agglutinated matter 
protects the contents from the effects of leaching thus making the 



14 



contents more nutritious than uncoated detritus. The excretory and 
feeding by-products deposited by the mussels are more densely 
colonized by bacteria than unfiltered material which further enhances 
the nutritional value of these products for detritivores since 
bacteria are a major nutritional component of the detritivore diet 
(Brinkhurst 1974). 

Much recent work has been done on filtration rates of the zebra 
mussel Dreissena polvmorpha and Asiatic clam Corbicula fluminea but 
there are few recent studies on the filtering rate of unionid mussels. 
The zebra mussel belongs to the superfamily Dreissenaceae and the 
Asiatic clam belongs to the superfamily Corbiculaceae. The mass of 
material sedimented by these mussels per unit area is greater than 
that of mussels in the superfamily Unionaceae because the 
Dreissenaceae and Corbiculaceae can occur at higher densities, but 
their ecological role as filter feeders is similar. 

Wisniewski (1990) measured a clearance rate of 4.2xl0~ 6 - 
5.0xl0" 6 pounds/hour for a 0.86 inch specimen of (L polvmorpha and a 
filtration rate of 5.3xl0~ 4 - 7.58xl0" 2 gal/mussel/h. He estimates a 
total of 0.017 lbs dry weight/ft 2 /day is sedimented by these mussels 
in lake Parteczyny. Izvekova and Lvova-Katchanova (1972) measured a 
filtration rate of 2.42xl0" 5 gal/h/lb for JL polvmorpha and estimated 
that the total population of zebra mussels in Uchinskoye reservoir 
sedimented 53.8 tons of dry matter/day. Reeders and Bij de Vaate 
(1990) estimate a filtration rate of 0.013-0.015 gal/mussel/hour. 



15 



Kryger and Riisgard (1988) find that the filtration rates of 
freshwater species are two to eight times lower than marine bivalves 
of comparable size. There is a decrease in the filtration rate with an 
increase in size. For example, Carver and Mallet (1990) measured a 
filtration rate of 6.98xl0" 4 - 1.45xl0" 3 gal/hour/lb dry weight for 
Mytilus edul is while Widdows et al . (1990) report filtration rates for 
the same species ranging from 8.9xl0" 4 gal/hour/lb for a 1.54xl0" 3 lb 
mussel and 3.9xl0" 3 gal/hour/lb for a 4xl0" 4 lb mussel. Kryger and 
Riisgard report .that , feeding rates of undisturbed freshwater bivalves 
can be four or more times higher than disturbed bivalves; for this 
reason they suggest that filtration rates may actually be much higher 
than has been previously reported. Data on the filtering rates of 
marine and freshwater mussels are summarized in Appendix C. 

Sedimentation of organic matter. Mussel beds may also cause 
planktonic material to settle by reducing current velocity through 
friction which disrupts the laminar flow of the water. The friction is 
caused by the rough surface of mussel beds, and as the current slows, 
planktonic material settles out of the water column (Holloway 1990). 
Haven and Morales-Alamo (1966) examined the biodeposition rate of 
oysters and estimated that seston filtered by the mussels is deposited 
as feces and pseudofeces about seven times faster than would occur by 
gravity. Similarly these authors cite Lund (1957) as calculating that 
oysters covering 1 acre of bottom could deposit 8.35 tons (dry weight) 
of fecal material in eleven days, equivalent to 12.86 lb/ftvyr. 



16 



Mussels process suspended organic matter and in this way make 
the material more available to detritivorous species. Obviously mussels 
cannot sediment more than what is available in the water column. 
Still, from this information one can appreciate the potential 
contribution to the detritus from planktonic material sedimented by 
mussels in the form of feces and pseudofeces. Bacteria comprise the 
dominant protein source for detritivorous fishes (Brinkhurst 1974). It 
is possible that in addition to providing other invertebrates with 
firm substrate and forage in the form of feces or pseudofeces, 
freshwater mussels may provide detritivorous fish with an enhanced 
food resource as well. 

Habitat permanence. The low production to biomass ratio (P:B) 
characteristic of unionid mussels means that a large population of 
long-lived adults is required to propagate the species. For the Thames 
river, Negus (1966) suggests a P:B ratio of 1:6 for mussels. Outside 
of man, adult mussels have few natural predators. The slow 
degradation of dead shells and long lives of adult mussels combine to 
make mussel beds a stable habitat which may allow for greater species 
diversity of other invertebrates which colonize mussel beds (Hargeby 
1990) 

Interstitial habitat. The interstitial space is important 
habitat for fish forage organisms (Cummins and Lauff 1969) and may be 
important for small fishes as well. Cobble substrate forms a 
temporally unstable refuge for small invertebrates in streams (DeMarch 
1976). As silt seasonally fills the interstices the animals are 



17 



eliminated from the habitat. Boulder (10 inches) substrate forms a 
physically more stable habitat because the particles do not move 
downstream and the interstices, which may be occupied by small 
invertebrates, crayfish, turtles or small fish, are not as readily 
filled by siltation. 

Immature insects appear to select substrate on the basis of 
particle size, or more precisely, interstitial space. When Cummins and 
Lauff (1969) offered invertebrates various sizes of particles 
associated with and not associated with silt, Caenis latipennis and 
Perlesta placida selected the interstices of coarse sediments. These 
authors state that substrate particle size and food supply are the 
primary macrodistributional influences on invertebrates. Egglishaw 
(1964) found a significant correlation between number of invertebrates 
and plant detritus in riffles and suggested that the animals were 
associating with, their food, rather than accumulating there solely for 
physical reasons, though Eriksen (1966) asserts that physical 
conditions such as water currents and oxygen are influential forces in 
the selection of crevice habitats. 

In laboratory studies with smallmouth bass, Sechnick et al . 
(1986) found substrate was important only when the fish could get into 
the interstices, while Rankin (1986) found that smallmouth bass in a 
natural stream select specific substrate types and suggests this is a 
response to prey distribution. 



18 



Substrate heterogeneity. Woodin (1978) suggests that substrate 
heterogeneity in the form of physical structure such as rock or cobble 
or biogenic structures such as the tubes built by polychaete worms can 
form refugia for benthic organisms. She cites the tubes of Diopatra 
cuprea , a marine polychaete, as an example of a biogenic refuge, one 
which is capable of renewing itself after destruction by a storm, a 
characteristic not possessed by refugia of physical origin. 

Mussel beds may perform a similar function in freshwater. Live 
mussels can maintain their position during both flood scour and silt 
deposition (Kranz 1974) and in doing so provide case-building insect 
larvae a solid place for attachment (Anderson and Vinikour 1984). 
Similarly the mussel itself could provide a current break for small 
benthic organisms (Hynes 1970). In a dense mussel bed the accumulation 
of dead shell material could provide attachment sites and interstices 
for non-bivalve invertebrates and spawning sites for lithophilous 
fishes (Pitlo 1989, Balon 1974). 

Bioturbation . Since mussels are living organisms they have the 
ability to adjust to minor perturbations in the environment (Kranz 
1974). Through their repositioning movements or other taxes mussels 
mix and turn the upper few centimeters of sediment as a plow turns a 
field. Through their movements, mussels held in an indoor tank with 
sand overlying gravel soon combine the two substrates into a 
homogeneous mixture (pers. obs.). In a mussel bed this action could 
bring shallowly-buried firm substrates to the surface creating 
attachment sites for other invertebrates and juvenile mussels 



19 



(Driscoll and Brandon 1973, Isley 1911) and help aerate the surface 
sediments for nitrogen and phosphorus cycles (Keeney 1973). 



20 



SUMMARY OF LITERATURE 

Habitat degradation threatens many species in Midwest rivers. In 
many cases maintenance of the habitat by human intervention carries 
prohibitive costs or is simply impractical. Mussel beds are a 
substrate suitable for foraging and spawning of fishes and can 
maintain themselves if not subjected to additional adverse human 
influences. 

Larval mussels complete metamorphosis and become dispersed 
through parasitism of a host fish. The question arises as to why the 
host fish occur in proximity to mussels at the time of glochidial 
release. The above discussion and references suggest that, as Coker 
and Howard suspected 75 years ago, forage occurring in mussel beds may 
attract the fish which are hosts for the mussels. 

In general glochidial hosts do not consume mussels as forage, 
but small cyprinids and percids which serve as forage for piscivorous 
fish do occur over mussel beds, probably in response to increased 
densities of non-bivalve invertebrate forage organisms attached to and 
feeding among the mussels. The mixture of mucus, bacteria, and organic 
particles found in mussel feces and pseudofeces provides nutritious 
forage for detritivorous invertebrates and is consumed by omnivorous 
fishes such as the fathead minnow ( Pimephales promelas ) as well (pers. 
obs.) . 



21 



It may be that the fish are responding to an area of increased 
forage rather than to the mussels themselves. Minnows, darters and 
young of larger species could find refuge from current or predators 
among the live and dead shells in mussel beds. These small fishes, 
together with non-bivalve macroinvertebrates might attract and provide 
forage for larger predacious fishes. 

The existing literature does not provide definitive answers to 
questions as to why fish occur over mussel beds, whether the fish 
benefit from this association, or whether the fish prefer mussels to 
other substrates. This study uses a laboratory approach to examine 
fish preference for mussels, seeks to identify what factors might 
attract fish to mussel beds and suggests how fish might benefit from 
their association with mussels. 



22 



RESEARCH QUESTIONS 

I. Do fish prefer mussels over other substrates? 
II. Is fish response to mussels based on the same factors as 
fish response to inorganic substrates? 

A. Is the response based on structure? 

1. Do the fish merely prefer to be over heterogeneous 
substrate? 

2. Can cobbles and mussels offer a refuge from predation? 

B. Is the response based on forage? 

a. Do the fish respond to invertebrates encrusting mussel 
shells and cobbles? 

b. Cah fish utilize mussel feces and pseudofeces? 

c. What is the nutritional content of glochidial packets 
consumed by fish? 



23 



APPROACH 

This study used three approaches in a laboratory setting to 
examine fish - mussel trophic relationships. The first was a substrate 
preference study, the second a predator-prey study and the third was a 
feeding study. 

Substrate preference trials determined whether the fish spent more 
time over mussels than over other types of substrate (cobbles and sand) 
and tested two hypotheses: (1) fish are attracted to the algae and 
small invertebrates attached to the mussel shells (the forage hypothe- 
sis), and (2) fish are attracted by the physical structure provided by 
the mussel shells (the structure hypothesis). The mussels were 
partially buried in sand in a natural position, with their pseudosiphons 
and the posterior portions of their shells above the sand surface. The 
cobbles were approximately the same size as the mussels and similarly 
placed in sand. Individual sport fish and individuals and schools of 
fathead minnows were given a choice of two substrates at a time in 
laboratory tanks. 

The forage hypothesis was tested by offering a choice between 
substrates colonized with attached algae and invertebrates and 
uncolonized substrates without attached organisms. The colonized 
substrates were cobbles and mussels from streams. Uncolonized 
substrates were mussels whose shells were scrubbed clean and cobbles 
obtained from roadsides or a quarry. The structure hypothesis was 
tested by giving fish choices between bare sand and uncolonized cobbles 



24 



or mussels. To check whether the shells alone were the attraction, 
rather than the living mussels, three species of fish (bluegill, white 
crappie, and juvenile channel catfish) were given a choice of empty 
mussel shells, as well as cobbles and live mussels. The empty shells 
were still joined in pairs by the elastic hinge ligaments, so they 
retained the external size and shape of the live mussels and were placed 
in the preference tanks in the same orientation as the live mussels. 
Empty shells were stored dry until placed in the preference tanks, and 
consequently there were no living organisms attached to them. 

The predator-prey trials measured the rate at which sauger 
captured fathead minnows over bare sand versus uncolonized cobbles. In 
each of six trials five fathead minnows were introduced on two 
consecutive days to a substrate preference tank containing a single 
sauger. 

The feeding studies stemmed from an observation during the 
preference trials of fathead minnows consuming mussel feces and a second 
observation of fish in the holding tanks eagerly consuming packets of 
glochidial larvae released by mussels. In the feeding trials, fish were 
held in aquaria alone or with mussels. The only food source provided 
during the eight two-week trials was from creek water continuously 
pumped into the aquaria. In every trial but one, the creek water was 
strained through a 0.01-inch mesh screen to remove most vegetation and 
invertebrates that were large enough for the fish to consume directly. 
The screens did not remove small particles that mussels are capable of 
filtering from the water. Thus the feces from the mussels were the 



25 



major source of food, and the utilization and nutritional value could be 
judged by comparing the weight change of fish held with mussels to fish 
held alone. In addition, the protein content of glochidial packets was 
measured to determine their potential nutritional value to fish. 



26 



PREFERENCE STUDY 

METHODS and MATERIALS 

Prior to the substrate preference trials we held the fish and 
mussels in separate systems. We kept fish indoors to acclimate them to 
confinement in tanks and to rid them of pathogenic agents such as 
monogenetic trematodes or Ichthvophthirius . We offered the fish live 
and formulated diets at a maintenance ration ad libitum while in the 
holding systems. 

• 
We held mussels in outdoor tanks or a nearby stream so that 
the shells would become colonized by non-bivalve invertebrates. We 
fertilized the outdoor tanks with F/2 algae food from Fritz 
Aquaculture Supply to encourage algal production and frequently added 
creek water to provide food for the mussels. 

Test chambers. The indoor experimental units offered a choice 
of two substrates, one placed on either side of the tank. We used two 
sizes of tanks. A set of six small tanks, 24 inches across the front 
by 48 inches on a side by 12 inches deep, and another set of three 
large tanks, 48 inches on a side by 12 inches deep. Baffles on the 
sides of each tank provided space for motor driven paddles which 
produced a current for maintenance of the mussels (Figure 1). Water 
circulated through the system via ports for water passage at both ends 
of the baffles. Paddles for each tank were driven by a 1/15 hp gear 



27 



66666666 -, 



Baffle 



mm§. 

/Substrate::: 
::: partition:;:;::^ 



Arrowsjlndlcat© 

direqtjonof 
.: : . : : : . :.watqr.flow . 



Camera 



Paddle- 
wheel 




APPLE PC 
with disk drive 



Figure 1. Top view of experimental tank and associated apparatus, a) 48 
inches in all tanks, b) 24 inches in small tanks, 48 inches in large 
tanks. 



28 



motor. The motors were connected to individual speed control switches 
so that uniform paddle velocity could be obtained in all tanks. 

Water Quality. Temperature in the experimental system was 
controlled by room temperature, addition of well water, and aquarium 
heaters. Water in each tank was pumped through a charcoal/zeolite 
canister filter to remove ammonia. The rotating paddles which created 
current for the mussels also aerated the water. Temperature and 
dissolved oxygen were measured daily, and ammonia and pH were 
monitored weekly in the holding and experimental units to ensure 
adequate water quality existed in all tanks. Automatic timers 
controlled the photoperiod and a partition eliminated entry of 
sunlight when the building door was opened; there were no windows. 

Fish Species. We used adult bluegill, sauger, white crappie, and 
fathead minnows; one-plus year old channel catfish, and young-of-the- 
year (YOY) largemouth bass, channel catfish and walleye; and tested 
each species individually. We used the small tanks with bluegill and 
one-plus year old channel catfish, and used the wider tanks with 
sauger, fathead minnow and all YOY fish; white crappie were tested in 
both sizes of tanks. 

We used single individuals of the large species and groups of 
ten fish with the small or young fish. Gorman (1988) found minnows in 
tanks formed cohesive schools when six or more individuals of a 
species were placed together. With the exception of bluegill and one- 
plus year old channel catfish there were six replicates of each 

29 



species. Two bluegill died from a parasitic infection; one catfish 
escaped. When we used the small tanks, six fish were tested at once. 
The large tanks took up more space so only three units were available 
at a time. 

Accl imation. The fish were allowed to acclimate in the 
experimental tank for at least one day and were not fed during data 
collection. Data collection began after the acclimation period. 

Control trials. A trial with the same substrate on both sides of 
the tank served as a control to determine fish preference for one side 
of the tank over another. Scrubbed mussels were the control substrate 
for bluegill, one-plus channel catfish and white crappie in the small 
tanks; sand was the control substrate for all other fish. The number 
of times the fish were located over each variable substrate was 
compared to the control substrate for each species. Table 3 lists the 
variable and control substrates used with each species. 

Variable Substrates. Live mussels colonized with periphyton and 
perizoon, rocks and live mussels which had been scrubbed clean, and 
sand were used in all trials. Rocks with perizoon were used with YOY 
fish and fathead minnows in the large tanks and clean, dead shells 
were used with bluegill, one-plus channel catfish and white crappie in 
the small tanks. Two to three inches of sand covered the bottom of the 
preference tanks. Substrate variables were placed on or in the sand 
and the sequence of variable presentation was randomized. We used 
several mussel species as the colonized or clean mussel substrate 

30 



Table 3. Variable and control substrates used with each species. 
Variable substrates are: a - colonized mussels, b - scrubbed mussels, 
c - sand, d - rock, e - colonized rock, f - dead shells. 



Species 


Control 


Variable 


Tank Size 




Substrate 


Substrates 




Bluegill 


Scrubbed mussels 


a,c,d,f 


Small 


White crappie 


Scrubbed mussels 


a,c,d,f 


Small 


H n 


Sand 


a,b,d 


Large 


Channel catfish 


Scrubbed mussels 


a,c,d,f 


Small 


Fathead minnow 


Sand 


a,b,d 


Large 


ti ii 


Sand 


a,b,d,e 


Large 


Sauger 


Sand 


a,b,d 


Large 


YOY Walleye 


Sand 


a,b,d,e 


Large 


YOY Channel catfish 


Sand 


a,b,d,e 


Large 


YOY Largemouth bass 


Sand 


a,b,d,e 


Large 



*) Small - 24 inch wide tanks; large - 48 inch wide tanks, 
YOY) Young-of-the-year fish. 



31 



during the preference trials depending on what was on hand at the 
time. Table 4 lists the mussel species used in the preference trials. 
Mussels were restricted to one side of the tank by a center barrier in 
the substrate which did not restrict fish movement (Figure 2). 

Data collection. Fish position data was collected with an image 
digitizing MicronEye camera linked to an Apple II GS computer. Back 
lighting of the tanks silhouetted the fish and allowed the camera to 
monitor the position of the fish. The tanks were viewed end-on as in 
Figure 2. Nocturnal readings were achieved by using a low light level 
with red light bulbs which produce primarily red and infrared light. 
The camera is sensitive to red and infrared wavelengths. Fish are 
relatively insensitive to these wavelengths when acclimated to low 
light conditions (Brett 1957). 

A software program (FishWatcher), developed by Robert F. II lyes , 
tracked the position of the center of the fish mass (a single fish or 
a school) and noted whether the center was over the right or left side 
of the tank (control or variable substrate). The position data were 
processed and recorded on a disk. At the end of data collection the 
percent of time the fish spent on the variable side of the tank was 
determined for statistical analysis. 

During a substrate trial data were collected for a minimum of 
one day and a maximum of three days. Upon termination of a given trial 
a new substrate was placed in the tank, and data collection began 

32 



Table 4. Mussel species used as substrate in the preference 
trials. Species at the top were used more often than those at the 
bottom of the 1 ist. 



Common Name 



Species 



Giant Floater 

Three-ridge 

Pocketbook 

Mapleleaf 

Washboard 

Yellow sandshell 

Pimpleback 

Pink heelspl itter 

White heelsplitter 

Fragile papershell 

Deertoe 

Hickorynut 

Rock-pocketbook 

Threehorn wartyback 



Anodonta qrandis 

Amblema plicata 

Lampsil is cardium 

Quadrula quadrula 

Megalonaias nervosa 

JLt teres 

G\. pustulosa 

Potamilus alatus 

Lasmigona complanata complanata 

Leptodea fraqil is 

Truncilla truncata 

Obovaria ol ivaria 

Arcidens confraqosus 

Obliquaria reflexa 



33 



Paddle 



Motor 



Variable 
Substrate 



Sand 



Baffle 



Figure 2. End-on view of substrate preference test tank. Variable 
substrate (rocks or mussels) is shown on the left side. Sand covers the 
bottom of the tank. Baffles account for the parallax of camera lens and 
keep fish in the camera's field of view. Water is circulated lb) 'motor 
driven paddles which move in the outer chamber created by the baffles. 



34 



again. This cycle continued until each fish was exposed to all 
substrates. Data collection began immediately after new substrates 
were placed in the tank. 

We randomized the order and side of substrate presentation with 
large fish but with the small fish ran the control trial first. The 
variable substrates were then always placed opposite the preferred 
side, though the order of variable presentation was still randomized. 

Data analysis. We used three-day substrate trials with bluegill, 
one-plus channel catfish, white crappie, sauger and the first set of 
fathead minnow trials. Analysis of the preference data indicated that 
some of the fish exhibited a preference on day one which diminished or 
changed to avoidance after three days perhaps in response to 
elimination of perizoon colonizing the surface of the substrate. For 
this reason, and to economize on time, data were collected for only 
one day with YOY walleye, channel catfish and largemouth bass and the 
second and third set of fathead minnow trials. We analyzed both one- 
day and three-day results of species which were used in the three-day 
trials. 

Analysis of species substrate preference is based on the value 
of the relationship of two measures: 

A = The percent of time the fish spent over the variable 
substrate. 

B = The percent of time the fish spent over the corresponding 
side of the tank in the control trial. 

35 



Relating the measures in either of two ways, illustrated as 'a' 
or 'b' below, determines how much increase or decrease in time spent 
over the variable substrate is possible compared to the control trial. 

a) (A - B) b) (A - B) 
(1 - B) B 

If A is greater than B, relationship 'a' indicates how much the 

time spent over the variable increased beyond the time that was spent 

over that side of the tank in the control trial relative to the 

increase in time possible for that side of the tank. The outcome 

measure is positive and ranges from near zero to positive one. If A is 

less than B relationship 'b' indicates how much less time was spent 

over the variable compared to what was spent in the control trial 

relative to the decrease in time possible for that side of the tank. 

The outcome measure of this relationship is negative ranging from near 

zero to negative one. 

We used a one-sample t test to compare the mean for each outcome 
measure to zero. If there was little or no difference between time 
spent over the variable versus the control the difference between the 
two percentages would be nearly zero. Conversely, if time spent over 
the variable was much more or less than time spent over the control 
the difference between the two percentages could be significantly 
different from zero. We used a treatment by time (3 or 4 X 2) repeated 
measures analysis of variance with a simple main effects test on 
treatment and time. 

36 



SUBSTRATE PREFERENCE RESULTS 

We use the term "prefer" to describe a relative increase in time 
spent over the variable compared to the control and the term "avoid" 
to refer to a decrease in that measure. Significant preference or 
avoidance indicates the measure differed significantly from zero (P < 
0.05). 

In general responses of the fish to a substrate were stronger 
the first day than after three days. Though the majority of responses 
by the fish to the substrates were not significant, the overall 
frequency of preference or avoidance responses to a substrate showed 
some interesting trends. 

Colonized mussels. Colonized mussels were preferred for at least 
part of the test period for all groups. The responses to this 
substrate were the strongest for all substrates tested, consequently 
this substrate elicited more significant responses than any other 
substrate (Figure 3) . 

Over the three-day trial period sauger exhibited a significant 
preference for colonized mussels during the day. Bluegill and white 
crappie in the small tanks also preferred colonized mussels though the 
responses were not significant (Figure 4). 



37 








en 



i- . •<- CO C >- 



C 00 



co § ^ cq <N ? 

6000 99 







<U «J C •" O QJ 






.c -o E «— J- 












1 1 




s_ > 01^ 






2=s .^s-° 


^ 













p 




X, n ' **^= 




j_j co •» 3 
a> v OJ 

2 > a. +»* ° § 

O « <1> •- 00 S- 
5, S 5 S- 0) C « r- 


>- 




i 
E 








•"" „ -r- +-> C 


i 
E 


!E 



£u **"•£« § . 

• £cS lea 

« g S- -F- -r- TJ — 1 


"■" 


2 


2 
E 




n 






3 ** -O •«- O « 0) 
00 ™ GO 4- r— 




>: 





2 






D 


"* JE c <o £ »-i 

00 ■- « -£ 1— ■ a> 
3 „ 01 f 1— c 


8 




SJ »4- -r- - O 




&•— C 0) O 1— 1 






•r- > .1- CO 










C 00 muj 


_i 




s> « :r 3 
1— > a> -* 2: 

T" i. .c ** 00 








52 +-> +j •-•<- 


CO 




+J O V) •>- r— I- H- 




Q. J3 3 ,_ Q) +j 


S 




d) 3 "T3 CJ) <TS 




oo • 00 00 E 3 O 




C """ t- 00 <0 




O •» 0) <1> V» r— 






00 S- +-> -U .2 ' £= 






O 




C£ Or— g-o nj 
a> 4-> 2-co .c 

co -<-> a> c h • » 


CD 




















d)U- C i- 00 c 


s 




s- a> ai -i- 3 




3 a.<+- +-> M- 




(J> >> 00 0J_ -r- +J >> 


9 




•r * J) 'r £ fl 




U_ T3 Q£ -O 2 O 1 



38 




O I 

g 

z 



n 



« .»-> _i _ m 



« c * ;* 

Q) CT>-t-> s- 



r- "> i— O) 



1 — o» «*« 



H *» >F- o 

p S 2 .. ■ 



>>+-> .t: ^ .- 

TJ o » « l/> 

2 * o — 



i en oj .c 

« air- o 



> 3 
3 O 



O GO 
CD (9 



Nl 



to o <_> 

V) o >- 

0) 



° = o.* ... 



v « 



</> 



+j d> a; , — 

W+, 01 A ^ 

8-S|«gg 
2 -£*§§ 



« 



O 
CO 



O 

CD 

d 



s 



o 

<N 
O 



9 



o 

CD 



en 

<U <U O -^ ,_ 

a> s- s- „, m 

S_ a) a) «• <D C71 

3 M- H_ +J _C C 

CD 0) 4- T- +-> 3 

•r- S_ -r- _£Z f0 O 

U_ Q."0 3 M- >, 



39 



Scrubbed mussels. Scrubbed mussels were not preferred as 
frequently as colonized mussels nor were the responses of the fish 
significant in as many cases (Figure 5). Over the three-day trial 
period there were no significant responses to scrubbed mussels from 
any of the fish tested (Figure 6). Though the bluegill differed in 
response to the substrate between the first and third day, in general 
the response of each species to the substrate after three days was 
similar to the response the first day of the trial. Again the response 
exhibited by the fish after three days was weaker than the response to 
the substrate the first day. 

Clean rocks. With the exception of white crappie in the large 
tanks, the groups of fish most frequently exhibited an avoidance 
response to the clean rocks. The white crappie preferred the clean 
rocks both day and night, while sauger and young channel catfish 
preferred the rocks only at night. None of the preferences were 
significant, but juvenile channel catfish and the second group of 
fathead minnows exhibited significant avoidances of clean rocks 
(Figure 7). 

The response of the fish to clean rocks varied over the three- 
day trial period. After three days all groups exhibited a preference 
for the clean rocks during day or night. Juvenile channel catfish 
continued to exhibit a significant avoidance of clean rocks during the 
day (Figure 8) . 



40 



J 





CQ 



LL S 

3: m 






n 



8 



o 
00 

6 



8 § 



o 
o o 



o 

9 



9 

9 



8 

9 



00 at c: 

a) 4-> •<- 



■S Mr- .„« 



il- 


cit 


+-> 










on 


JC 


0) 


<D 




r— 















a> 




Ob 






O. 

a. 














ro 




ro 






+-> 


0) 


s- 








c 



0) 


s- 
at 


u 


1/5 


<_> 








<4- 


a> 




CT 




>> 

ro 


c 


a. 


<4- 


•+-> 


J= 


c 
3 




■0 


3 


00 







S- 


a> 


.e 










O 


oi 




00 


(0 










j 
















3 




C_J 




00 








r— 


>- 




o> 






• •» 


at 










e 




c 






O 


01 




c 


at 




<v 









ro 


at 




a. 


c 




0>.C 


s- 




00 


ro 




QJ 


O-C 










3 




+-> 












Ol 






1 — 




^^ 


XJ 




"O 






> IT) 

ro O 


C3 
CD 


B 

a) 

> 


ro 
O 


00 

ro 

.a 


00 
Ol 

00 
to 

s 


a> 

3 

ro 

> 


V 
Q. 

O 


c 
at 

s- 

0) 


■■-> 

<_t 
<_t 


3-c 

+-> ■*-> 

3 

~ O 

O) E 

c a> 

O CT 


0) 




4- 






s_ 


■0 


> 

ro 
0) 


a> <4- 

N .,- 


c 

ro 
+-> 

at 


00 
ro 

S- 


ro 

CD 

C 
3 


00 


c 


+-> 


c 

ro 




3 



>> 






C 


(J 




c 


1 






at 






c 








i- u- 






33 






a> 


• r— 




s 2: 


at 


4- 


C 






1 


^ 


4- 


TO 


Q. 


TO 




c 








a. 


ro 







CD 


XI 


00 


ro 


at 


at 


Q.4- 






J- 


-C 


>, 


</» 


01 


>» a> 





+-) 


at 


at 


S- 




U 




ro 




cd 


Q.-4-> 


ro 


at 


M- 








c 




+-> 




ro 




a> 

+-> 

ro 


(0 


00 




1 


^ 


LT) 


O 


at 


3 


1 , 


en 


a> 


4- 


u 






rz 


s- 




-r— 


at 


1 


1 — 1 


3 


3 




C 


0. 









CJIXJ 


a> 


00 


1 


*» 


>i 




c 






_> 






Li- 


■»— 


00 


ro 


3 


1 — 1 


1 



41 






8 £ 

(D 



o 

CO 



d 



o 
o 



o 

9 



? 

9 



o 

CO 



« 3 1. C 
O r— «J O 
•«- ns i— 

«*- > I— 

■I- .£ Q.+J VI 

To J- s-° 



0) 



E 2 



% a) -§ -o § 

•^ 0) 3 C 

*• i = 

• Q. •« O 

■3*3 = ' 



v> 



J- .« 

5* 



S- <" <u 

£ =» a 

0) 



•+-> LU 



CO 



-Q <u o ,_ .e 

" t- •« CH- 

£■— -C nj 
O J3 •!— o <j 
4-> 3 0> 

V) 0) a) i— 

0> 3 ,_ oj 



(U 



-Q c c 

> -C 



a> 9 ts ■«-> 
Qi-^aa en 
• ■ c 



cu 



ns 



0"0 •» i 
3 CX-f- -* 

a> </> o e o 
■r- a> > ro o 
u_ i. nj -t-> >- 



42 





CO 



w 



o 



8 





- E tt) -O •- 




0) O -r- fO O) 




O S_ CL CD >> 




c <+- cc a> 




a> « 4-> i — 




S- +j i. (0 i— 




<1> e O U- <T3 




*4- 0J 3 




<u s- a) i 




v- <u +-> ai 




Q-U- T- _ c 




<+- -c rT 3 




a> -r- 5 rj o 




+-> T3 >> 




«J 1 




O >, ^ 1 








"O +-> O *~ ' LU 




C C 3 3 




•r- «J 




o ...•— ■ .- 




<A -r- j£ j; 




a> «♦- c z w 




3 •£ <oi ••- 




r- C+ju.11- 




ifl Dl -i-> 




> •■-i— .. « 




00 ,_ J. O 




a> *j J 




>5 E g»l- 




•i- +j M 5 o) 




*» J° rt <= 




■•" *: <u oo c 


1- 


H -M .r- « 


OKlQ, .C 


X 


Q. -O £ ' <J 


o 


3 « 


■ W S.O o> 


2 


(U u">C 




CO) 3 


_— . 


or (p ,»o 


H 


^ -^-JS.*: >> 




>> -i- oo 




<9 O SZ f 1 


I 


"O +■» x «+- 




Q 


O 00 o >- 




H-S(fl 


— l 


r- •« 


| 


co q-3 a, a) 




*4 53 = <" 




u « .. c s- 




pa:,- ro ^= 




*- ., i- j=-^> 




* ..- O 




C O) "O 00 




* • a> & c oo 




CD CD 3 n3 ro 



re 

O "O I 



dl or 



O tD 

0) > CO 
oo <w 

c 

O «'-. 

a. <v uo 

00 3 O 
0J i— • 

a; «j o 



o 

CO 



8 

d 



O 



o 



o 

9 



5 

9 



o 

CO 



1^ CD 

> a. 

0) -r- — 

i- -t-> 

3 « O 

en CD t- 

• r- QJ <u 

U.CN 



- E 
0) a) 

c en 

o s- 
«a 

w OO r- 

5 s - 3 
« ■»-> o 

*» >> 

CO O i 

en c 
s- c ca 



43 




•t— ™ 3» 



C l+_ _^ 



3 "r- o>^ 
*« a> J2 * X 



o 



■c <u.^ 



+-> 4-> 



^^ .. 



tfl _ -C ^ -^- 

>>£ * ?? w 



0) 



i/ii/i ai * 



-c </> - * * 



(0 a> _q 
u 

t/J C I 

*■ > . 

° > « 
QJ w 

a> > o 
= +j 5 

O J8 «s| 

Q. CH 

" £_ E 

ai c o 



O) 0) 

a> a> 



> 3 
3 O 






-^ ... 
C 0) 



+-> o 






CO CO 3j 

odd 



o o 

o 



o 

<^ 

9 



3 

9 



o 

CD 



a) ? 3 

..- o o 

Q = E 

<U Q> CD 



O) (DM- -i- 4J 3 

•i- S- ■«- SZ co O 



44 



Colonized rocks. Colonized rocks were tested with five groups of 
fish, two groups of fathead minnows, and all the young-of-the-year 
fishes (walleye, channel catfish and largemouth bass). Figure 9 
compares the responses to clean and colonized rocks exhibited by these 
five groups. The largemouth bass and two groups of fathead minnows 
exhibited the greatest differences in their responses to these 
substrates; these species preferred colonized rocks and avoided the 
clean rocks. The young channel catfish were mixed in their responses 
to these substrates and walleye avoided both substrates though the 
avoidance of clean rock was slightly stronger than the avoidance of 
colonized rock. 

Dead shells. Only three species were tested over the dead shell 
substrate: bluegill, white crappie in the small tanks and juvenile 
channel catfish. Figure 10 compares the response of these groups to 
this substrate the first and third days of the trial. Bluegill and 
channel catfish preferred this substrate both day and night for the 
duration of the trial. Conversely, white crappie exhibited a 
significant avoidance of this substrate both day and night through the 
third day. 



45 




03 




O 
CO 

6 



8 

d 



o 



o 
o 



o 



5 
9 



8 

9 



QJ 


S»l 


O) 


-t-> 




s- 


ro 


-1-J 


ro 


O 


c 






ro 


00 


■o 


o 


a> 



i— trt ro 



o> «-= ^ 


ro 


> ™4-> 




•1- 4j .,- O 




+JMS3 


-£_ 


•■-.O -t-> 




"> 3 l/> 




O (^ i_ </> 




Q. (U ,_ 


bs 


• •£.2 


P> 


-* fi 0> i. 


i- 


O ** i— 4-> 


ro 


o 


• 


*- o +-> s 


c 

3 

o 


+-> C O 


C 0) c 


rsus clea 

Response 

th differ 

thead min 




«„•'-« 




>* ȣ 


CI) 


"5 • « i 


>-> 


Q» 0, C 




N o E _ 


, 


'£ 5 = — • 


(T3 


£ "> •— HI 

O "O O 


3 


i— -i- o 


Oi 



o o " 

U iL .1—1 



I— I 3 



Si 






o 



> o <u •>- 



• o ■+-> ro ro 

Ol C C ur 

d) d) -i- u 

a> i- s- 4- 

s- a> ai •■- en 

3 m- 4- c c 

•r- S_ -.- -r- O 

U. Q.-O VI >» 



46 






CO 

H 
X 

_2 

2 
CO 



I 



I- 

X 

_2 

z 



I 



o n 



O 
CO 



8 9 



o 
o 



o 

CM 
9 



3 
9 



§ 
9 



<D <_■ O) 
> +-> -r- 

•i- (0 Q. 

+-> J- Q. 

• r-j_.ro 
</>(/>!- 

o __ o 



__ 

=» <_■ oo 

«" CO 

-k a) 

■S ^^ 

4-) _! 



00 



u-) 



</> 



O 2*° 

_ > v.^2 

O <4- 

IA > *- » <fl 

at •■" o 

2, *-■— 

>, o> ai a> 
«o <_ N c 
-o c c 

SaTg-5 

_= c *■ a> 

+_ a> __ 

i- +-> ^- 

■o o c c 

_>4- (U m 

«£fc§ 
£<^- 

•a"0 2_ ' 

i 0J TO 

a» +-» o 

c <« >><_> 

o o i— 

•i- +j .^ 

• -o c __ 

O C ro C 

—1 M- O « 



=5 3 C i— 
C0^— CO r0 

■i- f0 f- S 

LL > W lO 



47 



SUBSTRATE PREFERENCE DISCUSSION 
Overall the results from the preference trials indicate that the 
fish preferred mussels more frequently than clean rock. Table 5 
presents the preference results from day one of all trials. Colonized 
and scrubbed mussels were preferred more frequently than clean rocks, 
and no species exhibited a significant negative response to these two 
substrates but clean rocks were significantly avoided once both day 
and night. Though colonized rock was tested in only five trials, this 
substrate was preferred more frequently than clean rock and was never 
avoided. Most of the preferences and avoidances were not significant 
though fish most frequently displayed a significant preference for 
colonized mussels followed by scrubbed mussels and colonized rock. 
Clean rocks were avoided most frequently and dead shells, which were 
tested in only three trials, produced mixed results (Table 6). Figure 
11 illustrates the cumulative frequency of preference and avoidance 
during the first day of the substrate preference trials for each 
species. 

In the five trials with colonized rock, this substrate was never 
avoided and was preferred in several cases. In these same five trials, 
clean rocks were avoided by four species and the fifth species avoided 
clean rocks during the day. 



48 



Table 5. Summary of (day, night) results from the substrate 
preference trials for all species. + preference for the substrate, - 
avoidance of the substrate, = increase or decrease from the control 
was less than five percent. ( ) under scrubbed mussels indicates 
response to scrubbed mussels where sand was the variable, nt - not 
tested. 



Species 



Colonized 
Mussels 



Scrubbed 
Mussels 



Clean Colonized Dead 
Rock Rock Shells 



1 + 
w. 
w. 
w. 
w. 



Bluegill 

Bluegill 3 

1+ C. Catfish 
C. Catfish 3 
Crappie small 
Crappie small 3 
Crappie large 
Crappie large 3 

Sauger 

Sauger 3 

Fathead Minnow I 

Fathead Minnow I 3 

Fathead minnow II 

Fathead Minnow III 

YOY C. Catfish 

YOY Walleye 

YOY Largemouth Bass 



=»+ 


(-,-) 


+ » + 


nt 


+,+ 


+,+ 


(=»+) 


"*" 


nt 


+,+ 


- , - 


(+»-) 


~*' ~ 


nt 


+,+ 


-,+ 


(+.-) 




nt 


+ *+ 


+,+ 


(+»+) 


= ,- 


nt 


"*' 


+,= 


(+,+) 


- , - 


nt 


i 


" > — 


+,+ 


+ > + 


nt 


nt 


+*+ 


+,= 


+ > + 


nt 


nt 


+ *' + 


= ,= 


+ J + 


nt 


nt 


+ ,= 


+ ,+ 


-,+ 


nt 


nt 


+,- 


~ 5 


-,+ 


nt 


nt 


% = * 

+ ,+ 


+ »"* 


"' "* 


nt 

% + 
+ ,+ 


nt 

nt 
nt 


+ >% 


",+ 


",+ 


= ,+ 


nt 


+ * + * 


+ ,+ 


-,- 


— > — 


nt 


+ »+ 


",+ 


",- 


+ ,= 


nt 



*) Significant preference or avoidance, 
a) First day of three-day trial. 
YOY) Young-of-the-year fish. 



49 



Table 6. Summation of responses to the substrates for all 
species (from table 5). The first digit in each pair of numbers is the 
sum of the responses during the day for the indicated substrate. The 
second digit is the sum of the nocturnal responses. The paired digits 
in brackets represent the number of significant [day, night] 
responses. Equal means less than 5 percent difference between the 
response to the substrate and the control . 



Day 1 of 3-day trials 



Response 



Colonized 
Mussels 



Scrubbed 
Mussels 



Clean 
Rock 



Colonized 
Rock 



Dead 
Shells 



Prefer 


5,4 [1,0] 


5,3 [0,0] 


1,2 [0,0] 


not tested 2,2 [0,0] 


Avoid 


1,0 [0,0] 


0,2 [0,0] 


4,4 [1,0] 


1,1 [1,0] 


Equal 


0,2 


1,1 


1,0 


0,0 



Day 3 of 3-day trials 



Response 



Colonized 
Mussels 



Scrubbed 
Mussels 



Clean 
Rock 



Colonized 
Rock 



Dead 
Shells 



Prefer 


3,3 [1,0] 


3,2 [0,0] 


3,4 [0,0] 


not tested 2,2 [0,0] 


Avoid 


2,1 [0,0] 


1,3 [0,0] 


2,1 [1,0] 


1,1 [1,1] 


Equal 


1,2 


2,1 


1,1 


0,0 



Day 1 of all trials 



Response 



Colonized 
Mussels 



Scrubbed 
Mussels 



Clean 
Rock 



Colonized 
Rock 



Dead 
Shells 



Prefer 

Avoid 

Equal 



9,7 [3,3] 
1,0 [0,0] 
1,4 



7,8 [0,2] 
2,2 [0,0] 
2,1 



1,3 [0,0] 
9,8 [1,1] 
1,0 



2,3 [1,0] 
0,0 [0,0] 
3,2 



2,2 [0,0] 
1,1 [1,0] 
0,0 



50 



n 

E 

Z 




Colonized 
Mussels 



Scrubbed 
Mussels 



Rock 



Colonized 
Rock 



Dead 
Shells 



□ Total day M Total night 



Figure 11. Frequency of substrate preference for all species in the 
substrate preference trials. The numbers in parentheses above and below 
the bars indicate the number of significant preference and avoidance 
responses respectively for each substrate (P < 0.05). Responses less 
than 5% different from the control are not counted. 



51 



Colonized rock was never avoided by the five species tested with 
this substrate. Clean rocks were avoided by four of these five species 
both day and night and the fifth species avoided the clean rocks 
during the day. This suggests the fish were responding to forage or an 
odor associated with the colonized substrates. The clean rocks and 
dead shells were stored in a dry container prior to use in the 
substrate preference trials; these substrates were not colonized by 
bacteria or invertebrates before being placed in the test tanks. Since 
the rocks were not colonized by bacteria or invertebrates, the clean 
rocks may have had an unusual odor or lack of odor which caused the 
fish to avoid this substrate. Marzolf (1966) noted the amphipod 
Pontoporeia affinis showed significant selection of substrates whose 
surfaces had been "conditioned" by organic matter or bacteria. 



52 



PREDATOR - PREY TRIALS 
METHODS 

On two consecutive days, we placed five fathead minnows in each 
of three 48-inch-wide substrate preference tanks with a sauger. We 
performed all trials under daytime illumination with (overhead lamps 
on) sand on one side and rocks on the opposite side of the tank. 

We visually observed the fish until all the fathead minnows were 
consumed or for one hour, whichever occurred first. We generally 
recorded whether each fish was on the left or right side once each 
minute. The locations of the fish were recorded more frequently if the 
fish changed sides or there was a feeding event, less frequently if 
the fish did not move or there was no interaction. 

Analysis. We counted the number of minnows captured over each 
substrate, C s for sand and C r for rock and the time each species spent 
over the two substrates in the trials, T s for sand and T r for rock. 
For both species we multiplied the number of minutes for each time 
interval by the number of fish present on a given side of the tank for 
each time interval. This created a unit we call a fish-minute. There 
was only one sauger per tank so predator fish-minutes equal the total 
observation time. For fathead minnows we multiplied the number of fish 
remaining at each interval by the length of the interval and totaled 
the result. 



53 



The trials were of different lengths because the sauger consumed 
the fathead minnows more rapidly in some trials than in others. To 
give each trial equal weight in the analysis we calculated the mean 
trial length in fish-minutes for each species and divided this mean by 
the length of each trial. The mean duration of the trials for sauger 
was 35 fish-minutes; for fathead minnows the mean duration was 72 
fish-minutes. Division of these values by the duration of each trial 
for each species produces a weighting factor. This factor is less than 
one for trials longer than the mean, greater than one for trials 
shorter than the mean. When the weighting factor is multiplied by the 
fish-minutes spent over each substrate (T s or T r ) in a given trial, 
the result is the relative time the fish spent over the substrates in 
trials of equal duration (Appendix E). 

For each trial and species we also calculated the percent of the 
total fish-minutes spent over each substrate, %T S for sand or %T r for 
rock. Percent of captures over each substrate is represented by %C S 
for sand or %C r for rock. We analyzed both arcsin-transformed data 
and raw percent data; the transformed data produced similar results, 
therefore only the raw percent data are presented. 

We used a one-sample t test to examine whether the mean of the 
difference between the number of captures over the two substrates was 
significantly different from zero; the same test was used for fish- 
minutes. 



H :£(C = - C r j = and H Q : Z(l z - T_ r j = 
n n 



54 



We also used a one-sample t test to examine whether the mean 
percent of captures and mean percent time over each substrate were 
significantly different from 50 percent. 

H : %C S or %C r = 50 % and H Q : %T S or %T r = 50 % 

We calculated captures per fish-minute (C/T) for each species by 
substrate. This measure indicates the substrate-dependent minnow 
capture rate for the sauger and a substrate-dependent escape factor 
for the minnows. A high value indicates more minnows were captured per 
fish-minute. The values for each substrate were compared with ANOVA by 
species. 



55 



PREDATOR- PREY 
RESULTS AND DISCUSSION 

Sauger spent slightly more time over sand than over rock 
substrates. Fathead minnows spent twice as much time over rock than 
over sand. For both species the mean difference in fish-minutes spent 
over the two substrates was not significantly different from zero and 
the percent time spent over either substrate was not significantly 
different from 50 percent. Arcsin-transformed data produced similar 
results. 

Eighty-three percent of the minnow captures occurred over the 
sand substrate, significantly more than over rock. However, the mean 
difference between the number of captures over each substrate was not 
significantly different from zero and the percent of captures over 
each substrate did not differ significantly from 50 percent. The rate 
of minnow capture by sauger (captures/sauger-minute) and the fathead 
minnow mortality rate (captures/minnow-minute) did not differ 
significantly between the two substrates (Table 7). 

The minnows spent significantly more time over rock than over 
sand yet 83 percent of the total captures occurred over sand. The rate 
of minnow capture by sauger was approximately three times greater over 
sand than over rock, 0.26 and 0.09 respectively. The fathead minnows 
were twelve times more vulnerable over sand than over rock. Minnow 
mortality per minnow-minute was 0.25 over sand and 0.02 over rock, 
though the difference was not significant (P > 0.05). 

56 



Table 7. Mean time (T) spent over rocks and sand by sauger and 
fathead minnows, percent of minnow capture and capture per fish-minute 
by substrate in the predator-prey study. 

Sauger 



Substrate 


Mean 

Minnows 

Captured 


Percent 

Minnows 

Captured 


Fish-minu 
(T) 


ites 


Percent 

T 


Catch per 
Fish-minute 


Sand 


3.83 
* 


83.3 
* 


20.8 




59.5 





.26 


Rock 


0.83 


16.7 


14.2 




40.5 





.09 



Fathead Minnow 



Substrate 


Mean 

Minnows 

Captured 


Percent 

Minnows 

Captured 


Fish-minutes 
(T) 


Percent 
T 


Mortal ity 

per 
Fish-minute 


Sand 
Rock 


3.83 
* 

0.83 


83.3 
* 

16.7 


23.7 

* 

48.3 


32.9 
* 

67.1 


0.25 
0.02 



Significant difference between substrates (P < 0.05) 



57 



We compared the percent of time spent by each species over each 
substrate in the rock - sand preference trials with the percent time 
spent over the same substrates in the predator-prey trials using 
ANOVA. We used both arcsin-transformed and raw percent data for the 
analyses: since arcsin-transformed data produced the same result, 
untransformed data are presented. 

In the third day of the preference study, where behavior was 
compared to a sand-sand control, sauger exhibited a slight preference 
for rock in the rock-sand trial. Comparison of the percent time spent 
on either side of the tank within the rock-sand trial, as in the 
predator-prey study, indicates the fish spent a greater percent of the 
time over rock though the difference was not significant. There was no 
significant difference in time spent over a given substrate between 
the predator-prey and preference trials. 

The preference measure used in the substrate preference study 
indicated the fathead minnows frequently avoided clean rock. The ANOVA 
test of percent of time spent on each side of the tank in the sand - 
rock trial also indicated the fish spent more time over sand than 
rock, significantly more in trial three. In the predator-prey trials 
however, the minnows spent significantly more time over rock than over 
sand. Percent time spent by fathead minnows over a given substrate was 
significantly different between the predator-prey trials and trial 
three of the substrate preference study (Figure 12). 



58 



100 



80 



g 60 



O 

i— 


Q. 40 



20 




Predator I II III 

- Prey Preference 

Minnow 



Predator Preference 
- Prey 

Sauger 



Sand □ Rock 



Figure 12. Comparison of percent time spent over rock and sand by sauger 
and fathead minnow in the predator-prey study and the substrate 
preference trials. * Significant difference between the response to sand 
and rock within a trial. Different letters indicate a significant 
difference in response to a substrate between trials (P < 0.05). 
Unmarked columns within a substrate are not significantly different. 
(Differences in minnow response to sand were also significant, but 
opposite the response to rock). 



59 



Examination of the inverse of the capture rate provides some 
interesting information. The sauger spent an average of 3.8 minutes 
per minnow capture over sand and 11 minutes over rock. The fathead 
minnows spent 4 fish-minutes over sand for each one captured, but 
spent an average of 50 fish-minutes over rock before a minnow was 
captured. These results indicate that the irregular bottom profile 
formed by rocks in the experimental tanks provided the fathead minnows 
refuge from predatioh by the sauger. This finding suggests that the 
irregular bottom profile of a mussel bed could provide forage fish 
with some refuge from predation. 

These results, though limited in scope, agree with those of 
other predator-prey studies. Prey shift their distribution in response 
to a predator (Fraser and Cerri 1982, Stein 1979, Stein and Magnuson 
1976) and predators are more efficient in less complex habitats (Stein 
1977, Crowder and Cooper 1982). It seems reasonable that these 
predators could forage on the edge of a cobble habitat and thus be 
provided with both a better chance for prey capture over the open 
substrate as well as higher prey population densities over the cobble 
substrate. 



60 



FEEDING TRIALS 

MATERIALS AND METHODS 

During the substrate preference trials we observed fathead 
minnows consume mussel feces. To investigate whether fathead minnows 
may derive nutritional benefit from mussel feces or pseudofeces, we 
held fathead minnows in tanks alone or with mussels. The only food 
available to the fish during the eight, two-week trials was silt, 
phytoplankton and small invertebrates which entered the tank in water 
pumped from a nearby creek. 

Water Delivery. The feeding trials took place in 30 five-gallon 
aquaria contained in a 14 x 3.25 x 2.4-ft trough partially filled with 
water to form an isotemperate water bath. Water was pumped from nearby 
Quiver Creek with a 1/4-hp submersible pump replaced later by a 1/2-hp 
centrifugal pump. Creek water entered a head tank in the laboratory 
for aeration and mixing. Pump operation was regulated by a float 
switch in the head tank which maintained the water level within a set 
range therefore the head pressure was relatively stable even though 
the two pumps differed in rate of water delivery. An airstone and 
turbulence caused by water entering the tank kept sediment and 
planktonic matter mixed for delivery to the tanks. The water was 
delivered by gravity flow through a control valve for distribution to 
the aquaria (Figure 13). Each aquarium had a screened siphon, a glass 
lid to help prevent escape of test fish and an airstone (Figure 14). 
Water added to the aquarium exited through the siphons into the water 

61 




• >> o 2 

l/l -O +-> O 

'£ ,« > ° 

= J- 
■o = o> c 

2£ >> 

5 -° a» 
^ o. a> -*= 

g!| 2 

* CO -C 

,/. 2 *- ' 

"9 i_ OJ « 



c . ." 
■•" </> Jj 

mo™ 
a> o a> 

w> •<- -o 



o> 



a> a> 

si-** 

3 3* 
<_ = c 

<° «+^ 
a>-o 

41 c ,* 



a> rt P 



« * 

o 

O r— 



■o-o 



^o 
to (J 



> -r- 

■^ 3 s ° 

O ™ r~ 
a> s- o oi 
s- +j ^ s-r- 

3 O) +-> ro 

a> a) a) c i- 
■.- .c s- o s- 
u. t— <_> o +-> 



62 



Qlass Lid 




Figure 14. Detail of one of the thirty, five-gallon tanks used for the 
feeding trials. 



63 



bath in the surrounding trough. 

The method and volume of water delivery to the tanks varied with 
the trials. In trials one, two and three, two gallons of water was 
delivered with a 3/4-inch diameter hose to each tank two times daily. 
Prior to adding water to the aquaria we ran the pump continuously for 
15 minutes to ensure that water in the head tank was fresh and well 
mixed. 

For the first five days of trial four, water was delivered to 
each tank as in the previous trials. During the remaining nine days of 
trial four and the full duration of trials five through eight, the 
water was delivered through a manifold to all tanks in a constant 
stream at a mean rate of 339 gallons per tank per day. 

It was difficult to establish an equal flow to each tank with 
the manifold system and water flow varied significantly between tanks 
with highest and lowest flow rates. However, there was no significant 
difference in mean flow between treatments within a trial (P > 0.05). 

Water Treatment. In all trials except number four, the water was 
strained through a 0.01 inch-mesh net to remove most vegetation, 
zooplankton and insect larvae from the incoming water. In trial four 
the water was not strained. In trials one through three water was 
strained just prior to entry to the tanks at the end of the hose. In 
trials five through eight, when the manifold was used, water was 
strained prior to entry to the head tank. 



64 



Water Quality. We measured temperature, dissolved oxygen, pH, 
turbidity and total ammonia nitrogen daily in all tanks in trials 
three through eight. We used a Hach model FF-1A water quality test kit 
to measure ppm total ammonia nitrogen (+ 0.1 ppm) . We measured 
turbidity with a Hach DRL model 2506-05 water quality kit; results are 
reported in Formazin turbidity units (FTU) (+ 2.5 FTU). 
Milligrams/liter total ammonia nitrogen was measured with a Hach model 
FF-1A water quality test kit (+ 0.1 ppm). This kit was also used to 
measure pH in trials one through three(± 0.5). In trials four through 
eight, pH was measured with a Hanna pHep pocket pH meter (+ 0.2). The 
turbidimeter was obtained during the latter part of trial two, 
therefore turbidity was recorded for only a few days of this trial. 
Trial one was considered preliminary and no water quality data were 
collected. 

Treatments . We used four basic treatments for the feeding 
trials: blank, mussel, fish, and mussel -fish. Variations on the mussel 
and mussel-fish treatments used thick- or thin-shell mussels. Trials 
variously used one or two mussels per tank and one or two fish per 
tank. The blank treatment provided a measurement of background 
sediment and protein values for each trial. Similarly the fish and 
mussel treatments provided a means of comparing sediment weights and 
protein values with the mussel -fish treatment. 

Species. We used mapleleaf, three-ridge and floater mussels 
during the trials. Mapleleaf and three-ridge mussels are thick-shell 

65 



species, floaters are thin-shell mussels. Mapleleaf mussels were used 
in trial one only; three-ridge mussels were used in trials two and 
three, floaters were used in trials four through six and both three- 
ridge and floaters were used in trials seven and eight. We assumed 
that for a given wet weight, a thick-shell mussel would have less 
percent body tissue than a thin-shell mussel because more weight would 
be present as shell. Fathead minnow was the test fish in all trials. 
It was chosen for its availability and omnivorous food habits. 

Replicates . For each of trials one through six there were five 
empty tanks, five containing mussels, ten containing fish, and ten 
containing mussels and fish together. In trials seven and eight we 
used ten tanks containing fish, and twenty containing mussels and 
fish. Of the twenty mussel-fish treatment tanks half were fish with 
thick-shell mussels, the other ten contained fish with thin-shell 
mussels. Even though the tanks had lids, fish were occasionally able 
to escape. Tanks from which fish escaped or into which fish jumped 
were omitted from analysis as were tanks where fish or mussel 
mortalities occurred,' therefore sample sizes ranged from four to six 
in the trials. 

Treatments were randomly assigned to tanks in the system. 
Mussels were scrubbed clean then weighed to 2.2x10'^ lb in water prior 
to being placed in the tanks. Fish were weighed to 2.2xl0" 5 lb in 
water before being placed in the test tank. In trials one through four 
we randomly assigned fish and mussels to tanks within a treatment. In 
trials five through eight fish or mussels were assigned to tanks 



66 



within a treatment according to a predetermined mussel :fish weight 
ratio to match or alter the ratio used in previous trials. Table 8 
illustrates treatments and conditions used in each trial. 

End of Trial . At the end of a feeding trial the water was turned 
off, fish were netted out and weighed as described above. We dried the 
fish overnight in an oven at 167-185 °F and measured dry weight with 
an analytical balance. 

Since mussels were used in several non-consecutive trials, two 
conversion factors, one for thick-(three-ridge and mapleleaf) and one 
for thin-shell (floater) mussels were developed for calculation of dry 
tissue weight from wet, whole body weight. A sample of ten three-ridge 
and ten floater mussels was used to obtain the ratios of wet to dry 
weight. The mussels had been held indoors for 30 days, did not contain 
glochidia and appeared healthy. Each mussel was weighed wet, whole, 
then shucked and both wet body tissue and wet shell weight were 
recorded. Both shell and body tissue were dried overnight at 167-185 
°F. Dry weights of shell and tissue were measured. Shells were weighed 
to 2.2xl0" 4 lb with an electronic balance; tissue weights were 
measured to 2.2xl0~ 6 lb with an analytical balance. 

Appendix F gives the wet and dry weights of the mussels used for 
calculation of percent dry tissue weight from whole body wet weight. 
Dry weight is estimated to be 2.02 percent and 1.54 percent of wet 
weight for three-ridge and floater mussels respectively but these 
eans are not significantly different (P > 0.05). The conversion 



m 



67 



Table 8. Conditions and treatments for feeding trials. 

















Mean lb 






Mean 




Mussels/ 


Fish/ 




Mussel/ 
lb fish b 


Trial 


°F 


Flow 


Treatments 3 


Tank 


Tank 




1 


64.4 


2 


B,C,F,K 


1 


1 




0.035 


2 


74.3 


2 


B,C,F,K 


2 


1 




0.093 


3 


76.1 


2 


B,C,F,K^ 


2 


1 




0.088 


4 


74.1 


219 


B,M,F,N 


2 


1 




0.028 


5 


72.3 


339 


B,M,F,N 


2 


1 




0.031 


6 


63.1 


339 


B,M,F,N 


2 


2 




0.013 


7 


56.7 


339 


F,K,N 


1 


1 


0. 


.071, 0.016 c 


8 


51.4 


339 


F,K,N 


1 


1 


0. 


,064, 0.015 c 



a) B - blank, C - thick-shell mussel, F - Fish, K - thick-shell 
mussel-fish, M - thin-shell mussel, N - Thin-shell mussel-fish. 

b) Mussel :fish dry weight ratio. 

c) Thick- and thin-shell mussel/fish weight ratios respectively. 

Flow represents mean gallons/day water exchange per tank. Water 
was strained through a fine mesh net except as indicated by (*). Trial 
1 used mapleleaf mussels, the other trials with the thick-shell mussel 
treatment used three-ridge mussels. 



68 



factor for three-ridge mussels was applied to mapleleaf mussels. 

We siphoned the silt and water from each tank at the end of the 
trial and filtered it through an 3 . lxlO -3 inch-mesh plankton net. We 
assumed fecal protein, in the form of feces deposited by fish and 
mussels, would be associated with the sediments and would wash into 
the receptacle at the bottom of the net. The filtrate was either 
frozen and dried later or immediately dried overnight at 167-185 °F. 
We measured dry sediment weights with an analytical balance (+- 
2.2xl0~ 7 lb) then stored the samples in a dessicator for the sediment 
protein analysis. 

Sediment Protein. We used the micro-protein analysis procedure 
described by Smith et al . (1985) to determine if differences existed 
in the protein content of the sediment collected from different 
treatments. The assay, which measured protein rather than total 
nitrogen, was obtained from Pierce Chemical company. 

The proteins must be solubilized for the assay. We used a one 
percent solution of sodium lauryl sulfate (SDS), a detergent, to 
solubilize the proteins associated with the sediment. In preliminary 
extraction trials we mixed 0.013 gal of SDS with 0.001-0.002 lb of 
sediment and placed the mixture in an oven at 140 °F overnight. 
Subsequently, we found 85 percent of the overnight extraction level 
could be achieved using the same solvent-to-sample ratio for four 
hours at 140 °F. Since our interest was in relative concentrations of 
sediment protein between treatments rather than absolute 
concentrations, we considered this level of extraction to be 

69 



satisfactory. Protein values are expressed as a concentration of 
protein per sediment (ppt). 

Statistical analysis. We used the SAS PC general linear model 
(1985) analysis of variance procedure to determine if significant 
differences existed in fish weight change, percent fish weight change 
calculated as (final weight - initial weight)/(initial weight), 
sediment protein concentration and water quality parameters between 
treatments. We used the SAS regression analysis to identify 
significant factors in the variance of percent fish weight change. As 
in the previous studies we analyzed both arcsin-transformed percent 
data and raw percent data. The results were similar, therefore the 
untransformed data are presented. The level of significance for all 
tests is P < 0.05 unless otherwise noted. 



70 



RESULTS OF FEEDING TRIALS 

Fish weight change. Percent change in fish weight differed 
between treatments but was significantly different only in trials 
three and five (Figure 15). In trial three flow was two gallons per 
day. In this trial fish held with mussels lost less weight than fish 
held alone. In trial five, flow was 339 gallons per day; in this trial 
fish held with mussels lost weight while those held alone gained 
weight. Fish held with mussels in the other low flow trials (trials 
one and two) lost less weight than fish held alone but the differences 
were not statistically significant. In trials four and six, where flow 
was higher, fish held with mussels lost more or did not gain as much 
weight as fish held alone, though again the difference was not 
significant. In trials seven and eight where both thick and thin-shell 
mussel -fish treatments were used, the results were variable. 

In trials with thick-shell mussels, fish held with mussels lost 
less weight than fish held alone except in trial eight. Trials two and 
three were performed under similar experimental conditions and fish in 
the same treatments in these two trials exhibited similar weight 
changes. Mean fish weight change was significantly different between 
treatments in trial three (P < 0.05) but was not significantly 
different in trial two (P = 0.2062). Pooling the weight change results 
from these two trials shows fish held with mussels lost significantly 
less weight than fish held alone (P - 0.0187). 



71 







r— a. ^ OJ 



6«g5 
= **£ 

3 .,-.e </> 
SJ -o v> <» 

|. -i—3 

•>»''-■ 

— «•-•' 
•r- +-> C 

C w ai _ 
° 55^ * 

5 T- *» -C 

en «= c i 

■r- 'J- en 

a> JS i- 35 

3 +-> w 

+j * +j </> 
= ,„ Or- 






LO 



8 2 ° ? ^ $ ? 
e6uBL|3 m6]9M qsjj. jusojad 



</> i — 

■*-> ai 

1—1 c c .c 

a> a> oo 

Q) i. E i 

s_ oj +-> -^ 

U_ "O -M •*-» 



72 



Sediment protein. In general, the sediment protein 
concentrations measured for the mussel -fish treatment in a trial were 
equal to or greater than the sediment protein concentrations from the 
fish treatment. 

Sediment protein values tended to increase with higher water 
turnover (flow) in the tanks. Sediment protein varied significantly 
between treatments in trials one, three and five (Figure 16). 

The sediment protein concentrations in the low flow trials were 
higher in the treatments with mussels than in tanks which held no 
mussels. In these trials the fish weight change seems to reflect the 
increased protein in the sediment, though the differences were not 
significant. In the high flow trials sediment protein concentrations 
tended to be somewhat higher in tanks with mussels, though this was 
not always the case as can be seen in trials four and eight, and fish 
weight change did not appear to reflect the increased protein levels 
available in the sediment. 

In trial one, sediment protein concentration in the mussel 
treatment was higher than in all other treatments. In trial three the 
sediment protein concentration from the mussel treatment was 
significantly higher than treatments without mussels but not 
significantly greater than that measured for the mussel-fish 
treatment. The sediment protein concentration measured for the mussel - 
fish treatment in trial five was greater than the other three 
treatments but was significantly greater than the fish treatment. 



73 




.C CU >,"0 i — 

O +J r- 1— l— 

O) o C -= -= 

o c •-- «/> .E 



.= 4-> 2 

0) « ° C "O 

t/> c o •— 

+j C «*> a) i- 

O * ■"-> £H- 

s- c 

a w ai r , 

+J (l) 4-> -t- 
CL-t-J 0) 

•— <• (U i- I ••> 

I— +■> »— 

<o r j- a) fc 

*- S « «" 
^* * """ 

o •■ ' 

CO .,— 



c g-r 



a> 



» 8 



mom 

f» T- 

juauj|pas / uiajojd jdd 



e a) " 

•r- o c • •» 

■O a_ <U J^ 

QJ S- C 

"» «U » * 

<TJ ** -r- 1 

Q) -o ' 

•f- ^j CO 

<= .. 

<1> = **"- s- 

j_ -JJ -r- 0, 
3 <0 C l+_ 
CT> 0) U><4- 

•r- S- -r- -r- 
LL +J WD 



2 E 



74 



Figure 17 illustrates that over all trials and treatments fish 
weight change has a weak positive but significant relationship with 
sediment protein concentration {r c = 0.188). Note that Figure 17 
illustrates a separation between fish weight change in the low and 
high flow trials. This suggests that fish weight change was 
responding to changes in flow. When split by treatment, this same 
relationship holds for fish weight change in the fish only treatment 
(r 2 = 0.284) (Figure 18). The difference is significant (P < 0.05) in 
the thick-shell mussel-fish treatment (r 2 = 0.362) (Figure 19). The 
hypothesis that fish weight change responded to changes in flow is 
further supported by the results of the thin-shell mussel-fish 
treatments which were performed only at the high flow rate. Fish 
weight change from these treatments exhibits no significant 
relationship with sediment protein concentration and fish weight 
change (r 2 = 0.027) (Figure 20). Recall from figure 15 that sediment 
protein concentrations increased with flow. Sediment protein 
concentration exhibits a significant positive relationship with flow 
(r 2 = 0.5390) (Figure 21). 

We calculated 'net' sediment protein to more clearly define the 
relationship between the protein contained in mussel feces, 
pseudofeces or fish feces and fish weight change. Net protein is 
calculated as: 

Pn ij = p 9ij " Pb j 
Where, 

Pnij = Net Protein for an individual tank 'i' in trial 'j'. 

Pg^ j = Gross Protein for an individual tank 'i' in trial 'j'. 

Pbj = Mean mg protein/g sediment for the blank treatment in 
trial 'j', to account for background protein. 



75 




i + 



Q. 

"S. 
a. 



T3 



S- E 
0) •■- 
Q.-0 



° 2 

<U +J 

s- a. 
en Cl 
Q) "— ' 



eBump mB|9M qsy. jueojed 



O c 

•r- O 

u- o 



76 




If 

c 

i 

00 2 

is 

a- 5 

CO 3 






•r- -O 



O Q. 



o E 



efiueip m6|8M qsy. jauojsd 



0J 


« 


1- 


i- 


a> 


•t-> 


<D 


e 


Cf. 


CD 




O 




C 


CO 


o 
( ) 








c 


CD 




s- 


cd 


3 


+-> 


rr 


o 


•i" 


1- 



77 



OJ 



5 § 

d d 
ii ii 

u. a 



eo a 



£ 



s s ° ° ? 



_d ° 



3 
-o 



CM 


g 


M--o 


a) 

e 


.5? 


O "^ 

e 




X 


Q) •!- 


5 8 


+ 


en CD 
« a 




-e i- 


c 




u a. 


3 




■4-> ■»-> 



0»"0 

■i- c 



78 









- 




X 

+ ■ 

+ 
+ 


+ 
+ + 

+ + + 


- 


+ 
+ 


+ + + 

+ 

+ 
+ + 

+ + + " 


+ 
+ 

+ + 

+ 
+ 
+ + 


- 




+ 
+ 


+ + 


_ 






+ 


; 


0.0269 
0.2870 






- 


1 


1 


1 1 


i- 



Z--o 



eBueip in6|8M qsy. jueojed 



w ^ 




"- t/> 


•"I 




^> 


M 


£ 


&«> 


SI 


a 


go 


••^ 


^ 


■f? Q. 


c 


O) 


<J 


1 
•8 


I 




+ 


en. 


£ 




X o 


(0 




0) J- 

££ 

0) 0) 

a. o 

c 


■* 




t- o 
o o 

c c 

O v 
■f- <V 


CM 




(/> O 

0) s- 
s- a. 
o> 

a; 4-» 

a; c 

a> 


o 




• E 
o •«- 

C\J "O 
0) 

0) </> 

s- 



79 







+ + ■*■■& + + 


- 




+ 


+ + +t- jt 
+ ++ + +-H+ 


- 






+ +++► 44 +*+ + 








+ + 






+ 


+ 








+ + + 
+ + -H- ++ 


- 






+ 


- 


5390 
0001 






- 


o o 








L. Q. 








1 t . 




' - ++ - + J" ■!""■¥'■ 


■fHfr,4» r 



■<f 




+-> 


OJ 




cu 

e 

cu 

V) 


T— 




"V. 


N 




c 


CO 




cu 

+J 

o 
s- 

BL 


CO 
CO 




Q. 

a 

c 
o 


in 




+J 


(D 




(d 


OJ 




4-> 




>% 


C 




(0 







TO 


o 


OJ 


i 


c 
o 
o 


OJ 


^ 


c 




(0 




e3? 



OJ 



o3 



CO OJ CO 

)ueuj|pes / uiaiojd jdd 



CD,^ 



*2 

o c 
■a 

O o 

■•" X 
00 CO 

CT>j_> 

• >> 

C\J •!- 



•r- CO 

Li- £ 



80 



In trials seven and eight where there was no blank treatment we 
used the fish treatment as the background measurement. Calculation of 
these values provided a means of separating the effects of flow and 
sediment protein concentrations on percent fish weight change. 

When the results of all trials were pooled by treatment, the mean 
net sediment protein concentration of the thick-shell mussel treatment 
was significantly greater than all other treatments (Figure 22) and 
there were no significant differences between the other treatments. The 
mussel treatment in trial one was significantly greater than both the 
fish and mussel -fish treatments but in trial three was not significantly 
greater than the mussel -fish treatment. In trial five the net sediment 
protein in the thin-shell mussel-fish treatment was significantly 
greater than the fish treatment. 

The relationship between net sediment protein concentration and 
flow was not significant (r 2 = 0.008) (Figure 23) nor was the 
relationship between net sediment protein and percent fish weight 
change (r 2 = 0.010) (Figure 24). Water exchange rate (flow) accounted 
for thirty-two percent of the variance in fish weight change over all 
trials (Figure 25) . 

Comparison of the net sediment protein concentrations per gram of 
tissue for the three treatments (mussel, fish and mussel -fish) could 
show for example, whether thick-shell mussels produce higher sediment 
protein concentrations per gram dry tissue weight than thin-shell 
mussels. Figure 26 illustrates the results of dividing net sediment 
protein concentrations by total dry tissue weights for each treatment. 

81 



n 




w 


2 Q 


CVJ 


2 to 


+ 


5 CNI 

1 





■ + 

L 







-C oo >■> i oo 


2 




O d) .— -Si s- 




CO O -t-> O d) 


i£ 


oo 


0> C C •<- -O 

CD ro _C E 
. S- (_>■(-> 3 


Li. 






** H- -r- 4_> • 






•r- C .,- 01 






^-^■O Dljr- 


2 




c ■!-> «" -a «" 


* 


h- 


a> C r- <* 


LL 




sed 
nifi 
re n 

sh 
ell 










z 




^».,- ^ 00 
C v> w ( 1 


LL 


CO 


Is!*! 






:> 




pro 
dica 

col 
ne; 
th t 


z 






11, 


in 




2 





2 

LL CO 
O 



OJ 



s 



©.in o in 

lueiuipes / LBiojd idd jefs| 



o 




o. 




, — 




c 


Q 




r^-" 


CU 
CO 










00 




0) 

-t-> 
o 


01 


fO 

s- 


01 
00 

3 

F 


3 




Q. S- 


CO 

e 


0) 


CU 

J= 

01 

I 

c 




F 




■4-> 








•o 

01 


CU 
CU 
14- 


'5 


(J 


JCZ 






O) 


c 


+-> 


f 




-t-> 


-C 


CI) 








cu 


-l-> 


F 




■ * 




z 




+■> 




00 


fd 




c 


CO 


<_) 


r^ 


•i— 






CU 




cu 


i- 






i- 




00 


+-> 


CM 


-M 


+-> 


4-> 


00 




C 




c 


3 


O 




O) 


c 


o> 


f= 


-(-> 


<1> 


f= 


CU 


s- 






s- 


+-> 


o> 


cu 




i- 


3 


ro 


S <*- 




O) 


CT> (!) 


+-> 


u- 


O) l*_ 




i- 


CU 




-C 


o> 


Lu +-> -Q TO 


00 


s_ 



82 



£fc 



Pttl 



■¥ 

K V 

+ ++ 

+ + +# 



+ 

+ + -H- 



CO 

o 


1 


8 


o 


o 


cm" 


II 

Q. 



CO 







■H 






«J 






t- 






4-> 






c 






Q) 






O 






c 




|> 


o 




■o 


c 




■■■** 




CM 


g 


0) . 


OJ 


5 


? >» 

J- co 




CO 


Q.-0 




O 


4-> VI 

c c 


Cft 

in 




CD o 

Si— 



$ 



8 



8 



tiii ,++, iii i , i iiii|i i i^ i if i i ,-w-, , rt « 

CO CM 00 <«t O "»f Cp 

jueiuipes / U|8)0Jd ;dd }9n 



o 


i~ 


01 

Of 

t- 


0) 

5 


cr 




cu 
ex. 






rtj 




-o 


CM 


c 




ro 


QJ 


cu 


s- 


e 


=3 




C7>-0 


••" 


c 



83 



+ 



w 



+ 



r*t* 



a a 



+ 

a m 
+ + + a a 

£+ " ffl B 

+ + + ¥ "b 

* + 
+ + 



+ + + 

+ 



C\J 



O 06 

*■* co 

o c\j 

d d 

II II 

b. Q. 



-H-%* 



H + 



a 

+ 



- C\J 



o 

CO 



a 



e5uBL|0 }q6|6M qsi; juaojsd 



8 



£ 



1 + 



c 

'1 

a 

z i 



en 



0) a> 

IS 

o ■■- 



a> 

a: c 

o 

CM ™ 



84 



+ 



+ :* + 

-H-H- + 
+ +-h.*HH|.+ + 



+ + 

+ \ 



05 

r^ 

CNJ 

CO 

d 



OJ 



II II 



8 8 



+ + 



7- >% 

g 

•a 



o S 



86UBLJD }LjBj8M Ljsy. jueojQd 





T7 


LO 


\ 




to 




c 


QJ 


o 


t- 




3 


1 — 


C7> <TJ 


••— 


en 



85 



- x: aj 




§ S ° S § 

do d d 9 

}Lj6ieM Ajp |b;o; q| / ftuQWipss / uiejcud }d<j) 



>i"0 +-> -o 


1 


i_ 


0J 


o 






■o 


i- 


C 


a> 


01 




01 




-c 


J= 


•o 


4- 


0) 




00 


E 




S- J= 




3 




ro 


00 


c 



Q. -13 r- 14- -C 



:-m*S 



+-> 


(TJ 


(T3 




2 


c 


> 


B 


{- 




Q> 






O 


"O 


E 




-C 




^ 












-o 








-E 


01 


rC 


3 


■o 


.e 






00 


ru 




\ s- 


B 


-C 




c 


+-> 


F 




«4- 


• f— 




3 


-E 














+-> 


o- 


O 




1 


o 


e 


U l*_ 












2 


CL-c 


■o 


1 




■»-> 


a> 






... 


&*■ 


E 




0) 
00 


e 


jE 


=3 


00 


3 
E 




4J 




00 




-t-> 




^ 


3 

E 


,_ 


f0 


c m 






•*- 


O 


1— 


a) 

-E 


c 




O 


ai 


00 


a> 4-> 






o 


B 


V 


B 




0) 








o 
<-> 




.e 




<T» 






00 




a> 


X.E 




QJ 

O 

S- 


t- 


•p- 


4-> 


ro 


4J 








e 


ro O 




u 






00 O 


. % 


n» 


1~ 


• l 


— +J 



& E <U 00 Q) 

£ s- •■- S- = <+- 

^r o a» E Q> 

W E — "— oo 



2 °V?^ 



>»>» 



00 



10 s « « <-> z 

CM O <j£ • 



00 



00 



O 00 CT> CT14-' 00 
•e- ..- -i- .,- -r- 3 

u_ +-> oo oo 2 e 



86 



Looking at all trials, there was no significant difference in net 
sediment protein produced per pound dry tissue in different mussel 
treatments and no significant relationship between these two variables 
(r^ = 0.001). In trial six the net protein/lb tissue was significantly 
greater in the fish treatment than in the mussel -fish treatment. In 
general, the mussel -fish treatments produced higher gross (and net) 
sediment protein concentrations, but the increase was apparently due to 
the increased body tissue mass in the tank rather than an enhanced 
protein producing property of the mussels themselves through the 
production of pseudofeces. 

Sediment weight. The total amount of dry sediment collected from 
the tanks at the end of the trials varied significantly between 
treatments in six of the eight trials. In the low flow trials the 
tanks containing mussels had more sediment than tanks without mussels, 
significantly more in trials two and three. In trials four, five and 
six, tanks containing fish had significantly less sediment in them 
than blank or mussel treatment tanks. In trial eight the fish 
treatment tanks contained significantly more sediment than the thick- 
shell mussel -fish treatment tanks (Figure 27). 

Water Qual i tv . Dissolved oxygen did not differ between 
treatments within a trial (Figure 28). Turbidity varied significantly 
between treatments in trials two and three where flow (turnover) was 
low (Figure 29). In trials two and three, mean turbidities for the 
blank and fish treatments were significantly higher than the mussel 



87 




Figure 27. Mean dry sediment weight collected from each treatment in the 
feeding trials. Significant differences within a trial are indicated by 
different letters (P < 0.05). Unmarked columns within a trial are not 
significantly different. B - blank; C - thick-shell mussel; F - fish 
held alone; K - fish held with thick-shell mussels; M - thin-shell 
mussel; N - fish held with thin-shell mussels. Numbers refer to trials. 



88 




BCFK BCFK BMFN BMFN BMFN KFN 
2 3 4 5 6 7 

TRIAL 
TEMP% SATURATION 



KFN 
8 



Figure 28. Concentration (ppm) and percent saturation of dissolved 
oxygen for each treatment in the feeding trials. No significant 
differences between treatments; no data were collected for trial one. B 
- blank; C - thick-shell mussel; F - fish held alone; K - fish held with 
thick-shell mussels; M - thin-shell mussel; N - fish held with thin- 
shell mussels. Numbers refer to trials. 



89 



100 




BCFK BCFK BMFN BMFN BMFN KFN KFN 

2 3 4 5 6 7 8 

TRIAL 



Figure 29. Mean turbidity in formazin turbidity units (FTU) for each 
treatment in the feeding trials. Significant differences within a trial 
are indicated by different letters (P < 0.05); unmarked columns are not 
significantly different. No data were collected for trial one. B - 
blank; C - thick-shell mussel; F - fish held alone; K - fish held with 
thick-shell mussels; M - thin-shell mussel; N - fish held with thin- 
shell mussels. Numbers refer to trials. 



90 



and mussel-fish turbidities. This same pattern of turbidity 
differences between treatments continued through the other trials but 
the differences were not significant. 

Figure 30 illustrates mean un-ionized ammonia concentrations, 
corrected for pH and temperature. These values did not vary 
significantly among treatments except in trial four in which mean un- 
ionized ammonia concentration in the mussel and mussel-fish treatments 
was significantly higher than that of the blank and fish treatments. 

Un-ionized ammonia concentrations greater than 0.125 ppm causes 
reduced growth in channel catfish (Robinette 1976, Stickney 1979) and 
levels as low as 0.0125 ppm may adversely affect the growth of trout 
(Piper et al . 1986). The concentration of un-ionized ammonia exceeded 
0.125 ppm in trials two, three and four and may have affected the 
weight change of fish in trial four. During trial four, mean ammonia 
concentration was lower in tanks where fish were held alone. Fish in 
this treatment did not lose as much weight as fish held with mussels 
though the difference was not significant. Elevated ammonia readings 
were also recorded in trials two and three; the ammonia levels in 
trial three were significantly higher than in trials five through 
eight which were the high flow trials. In trials two and three there 
was no significant difference in un-ionized ammonia concentration 
between treatments and fish held alone lost more weight than fish held 
with mussels. 



91 




BCFK BCFK BM FN BMFN BMFN KFN KFN 

2 3 4 5 6 7 8 

TRIAL 

pH TEMPO 



Figure 30. Mean un-ionized arranonia concentrations for each treatment in 
the feeding trials. Values are temperature and pH corrected. Significant 
differences within a trial are indicated by different letters (P < 
0.05). Unmarked columns are not significantly different. No data were 
collected for trial one. B - blank; C - thick-shell mussel; F - fish 
held alone; K - fish held with thick-shell mussels; M - thin-shell 
mussel; N - fish held with thin-shell mussels. Numbers refer to trials. 



92 



Mussel weights. We weighed the mussels before and after each trial, 
recorded the difference in weight and calculated a percent weight 
change for each mussel. These data are somewhat dubious though because 
mussels retain various volumes of water in their mantle cavity depending 
upon variables such as how the mussel was handled, rate of valve closure 
and reproductive state. Recall from Appendix F that water in the mantle 
cavity ranged from 16 percent body weight in thick-shell mussels to 50 
percent in thin-shell mussels. The variation in how much of this water 
is expelled as the valves close causes difficulty in measuring mussel 
weights even when the mussels are weighed in water. Mussel weight change 
is briefly discussed below but one must bear in mind that weighing error 
may be significant and is difficult to avoid even with careful handling. 

Mussel weight change and percent weight change were 
significantly different between treatments only in trials five and 
seven. Arcsin-transformation of percent weight change produced the same 
results as percent weight change. In trial five mussels held with fish 
gained 3.82 percent body weight (0.0083 lb) while mussels held alone 
lost 1.33 percent body weight (0.0069 lb). In trial seven the thin-shell 
mussels held with fish lost 1.98 percent body weight (0.0048 lb) and 
thick-shell mussels held with fish gained 0.23 percent body weight 
(0.0017 lb). There were no significant differences in mussel weight 
change in the other trials and change in mussel weight showed no 
consistent relationship with change in fish weight. 

Table 9 lists correlation coefficients for several biological and 
environmental parameters with percent fish weight change. All factors 
were considered separately for correlation and together in a stepwise 
multiple regression model. 

93 



Table 9. Correlation and regression coefficients of biological 
and environmental variables with percent change in fish weight. Sample 
sizes differ due to uncollected data or absence of mussels in 
treatments. 



Parameter 


r 


r 2 


N 


Biological 








Sediment Protein 


0.4344 


0.1887 


155 


Net Sediment Protein 


0.1006 3 


0.0101 a 


139 


Dry Mussel Weight 


-0.2153 


0.0463 


87 


Dry Fish Weight 


0.5159 


0.2661 


155 


Mussel Weight Change 


0.0248 3 


0.0006 a 


87 


Environmental 








Liters/day water 


0.5726 


0.3279 


155 


Temperature 


-0.3299 


0.1088 


155 


Dissolved Oxygen 


0.2046 


0.0419 


155 


pH 


-0.5233 


0.2738 


139 


Unionized Ammonia 


-0.5880 


0.3457 


139 


Turbidity 


-0.1476 3 


0.0218 3 


134 


Sediment weight 


0.4011 


0.1609 


155 



(a) - No significant correlation (P > 0.05). Unmarked coefficients are 
significant. 



Net sediment protein = total sediment protein - mean sediment 
protein of the blank treatment in a given trial. If no blank was 
present (trials 7 and 8) the mean sediment protein of the fish 
treatment in that trial was used. 



94 



Sediment protein, sediment weight, dry weight of fish, gallons per 
day water exchange and concentration of dissolved oxygen exhibited 
significant positive relationships with percent fish weight change. Dry 
weight of mussels, pH, un-ionized ammonia concentration and temperature 
exhibited significant negative relationships with percent fish weight 
change. 

The stepwise regression model indicated that dry weight of fish, 
gallons/day water exchange, temperature and sediment weight accounted 
for 59.23 percent of the variance in fish weight change for all tanks 
which contained fish (fish, and mussel-fish treatments). Dry weight of 
fish, sediment protein concentration and net sediment protein 
concentration accounted for 40.00 percent of the variance of fish weight 
change in the mussel -fish treatments. Heavier fish lost less weight than 
smaller fish. Dry weight of mussels was not a significant effect in the 
multiple regression analysis. 

Temperature and pH interact with ammonia and affect the percent of 
ammonia which occurs in the un-ionized (toxic) form. Separate analysis 
of the high and low flow trials indicated un-ionized ammonia 
concentration was negatively correlated with percent fish weight change 
in the low flow trials. In the low flow trials warm temperatures and 
high pH together with low daily water exchange may have interacted with 
high total ammonia concentrations and adversely affected the weight 
change of fish in these trials. Regression analysis of the low flow 
trials indicated that temperature, pH, and un-ionized ammonia, and 
dissolved oxygen concentrations were colinear and could not be analyzed 



95 



as separate effects. 

In the high flow trials water in the tanks was exchanged at a mean 
rate of three times per hour, a rate great enough to flush out un- 
ionized ammonia before it accumulated to toxic levels. The apparent 
positive relationship of percent fish weight change and un-ionized 
ammonia concentration may be a result of falling temperatures 
accompanied by lower concentrations of un-ionized ammonia and decreasing 
percent fish weight change. 



96 



FEEDING TRIALS DISCUSSION 

There was little evidence from the feeding trials to indicate 
that mussel feces are beneficial as forage for fathead minnows. Fish 
held with mussels exhibited less weight loss than fish held alone in 
three of the eight trials but never gained more weight than fish held 
alone. Significant differences existed between treatments in trials 
three and five. In trial three fish held with mussels lost less weight 
than fish held alone. In trial five fish held with mussels lost weight 
while fish held alone gained weight. There were several factors 
interacting within the series of trials which apparently influenced 
fish weight change to various degrees over the course of the tests. 

Fish Held Alone. Weight change of fish held alone appeared to 
reflect changes in flow and temperature. The following discussion 
refers to Figure 31 which illustrates fish weight change, mean 
temperature, mean flow and mean mussel weight per fish in each trial. 

In trial one mean temperature was 64.4 °F, this increased to the 
high 60's for trials two through five, then dropped to 62.6 °F in 
trial six, 55.4 °F in trial seven to a low of 50.0 °F in trial eight. 
In trials one through three water exchange was two gallons per day, in 
trial four water exchange increased to about 220 gallons per day 



97 



30 



a, 20 

j? 10 
u 



O) 

1 



§-20 

is 

Q. 

-30 



-40 



11 



II 



EL^kl 



+ 2 STD DEVIATIONS 
MEAN WEIGHT CHANGE 
- 2 STD DEVIATIONS 



0.13g= 

s 

0.09 § 

2. 

0.04 5 

y 



FK FK FK FN 
12 3 4 

Thick-shell mussel 



FN FN FKN FKN 
5 6 7 8 

S Thin-shell mussel 




Temperature 



ffl Gallons /day 



Figure 31. Percent fish weight change in the feeding trials compared 
with mussel -to-fish weight ratio, temperature and daily water exchange 
(gallons/day). Significant differences in percent fish weight change 
between treatments within a trial are indicated by different letters (P 
< 0.05), unmarked treatments are not significantly different. B - blank; 
C - thick-shell mussel; F - fish held alone; K - fish held with thick- 
shell mussels; M - thin-shell mussel; N - fish held with thin-shell 
mussels. Numbers refer to trials. 



98 



and reached a maximum stable value for the remainder of the trials at 
339 gallons per day. The low flow was intended to limit external 
sources of forage so that the minnows would have relatively little to 
consume except for mussel feces. 

Weight change of fish held alone appeared to reflect changes in 
flow and temperature, more so than the weight change of fish held with 
mussels. Fish held alone lost five percent body weight in trial one 
where flow was two gallons per day and temperature was 64.4 °F. In the 
next two trials flow was maintained at two gallons per day, but 
temperature increased to 73.4 and 75.2 °F. In these two trials fish 
held alone lost about eleven percent body weight, perhaps in response 
to increased metabolic demand created by high temperature and the 
extremely limited forage present in the low volume of daily water 
exchange. Temperature remained fairly steady through trial four at 
73.4 °F but mean flow was increased to about 220 gallons per day. This 
increased the forage available to the fish which may be reflected in 
the low percent weight loss of the fish held alone in trial four. This 
pattern continued into trial five where temperature remained at 71.6 
°F and flow increased to 339 gallons per day. In trial five fish held 
alone gained eight percent body weight, though the sample size was 
only seven due to escapes. 

In trial six mean temperature decreased to 62.6 °F. Mean percent 
weight change decreased slightly as well, though the decrease in 
weight change cannot be attributed to temperature alone since there 
were two fish held per tank in trial six. 

99 



In trial seven, temperature continued to decrease to a mean of 
55.4 °F. Fish held alone lost three percent body weight which appeared 
to reflect the decrease in temperature. In trial eight, mean 
temperature was 50.0 °F; fish held alone in this trial gained two 
percent body weight. Two factors might have affected the results of 
trials seven and eight: temperature fluctuation and reduction of 
treatment sample size due to mortalities. The temperature in trial 
seven dropped from 64.4 to 46.4 °F over a period of four days in the 
middle of the trial, then rose again to 55.4 °F by the end of the 
trial. Though the mean temperature was low in trial eight 50.0 °F, 
temperature remained fairly stable ranging from 48.2 to 53.6 °F over 
the course of the trial. Four fish were lost to mortality in the fish 
alone treatment in trial eight; the mortalities were not size 
selective. The low sample size and moderate temperature fluctuation in 
trial eight may have resulted in the apparent increase in mean weight 
of fish held alone, but change in fish weight was not significantly 
different between trials seven and eight. 

Fish Held With Mussels. The mean weight change of fish held with 
mussels appeared to vary with temperature and flow but was also 
affected by the presence of mussels in the tank. In trial one, fish 
were held with a variety of sizes of mapleleaf mussels with mussel 
tissue/fish ratios ranging from 0.013 lb mussel/lb fish to 0.138 lb 
mussel/lb fish, this wide range of mussel sizes may be reflected in 
the large standard deviation of the mean weight change for fish held 
with mussels in trial one. In trials two, three and four weight loss 
of fish held with mussels was about seven percent body weight, even 

100 



though flow increased 100-fold and the weight of mussel per fish 
decreased by about 60 percent in trial four. 

In trials two and three fish held with mussels lost less weight 
than fish held alone. In these trials forage was very limited and 
temperature was 73.4 to 75.2 °F. In trial four, where flow (and 
forage) increased, weight change of fish held with mussels was less 
than fish held alone. Perhaps at the very low forage levels in trials 
two and three the fish were able to take advantage of mussel feces, 
since mussels can filter and coalesce particles which would otherwise 
be too small for the fish to consume. But when other sources of forage 
were available the mussels may have competed to a degree with the 
fish. 

The results of trial five are similar to trial four. Fish alone 
gained weight while fish held with mussels lost weight. 

The results of trials six are difficult to interpret since there 
were two fish per tank which decreased the net weight of mussel per 
fish and temperature decreased. Fish held alone gained slightly less 
weight than did fish in trial five, while fish held with mussels 
gained the greatest mean weight of any mussel -fish treatment. 

Mean turbidity in trial six was highest among the high flow 
trials. Mussels do not produce pseudofeces until suspended particle 
concentrations surpass what is required for maintenance (Winter 1978). 
Perhaps the higher turbidity present in trial six allowed the mussels 

101 



to produce pseudofeces, and the fish held with the mussels were able 
to consume these materials and thus exhibit less weight loss or 
greater weight gain than fish held with mussels in the previous 
trials. The filtering action of the mussels in the small tanks may 
still have competed with the fish for small food items resulting in 
less weight increase than that exhibited by fish held alone. 

It is difficult to assess what might have occurred in trials 
seven and eight. Perhaps the presence of mussels in the tank 
maintained a constant source of forage and thus allowed fish held with 
thick-shell mussels to maintain their weight, while fish held alone 
lost weight. The fish in the thick-shell mussel treatment were held 
with 0.070 lb mussel per lb fish, while the thin-shell treatment had 
only about 0.015 lb mussel per lb fish. It is likely that 
proportionately more feces and pseudofeces were present in the tanks 
with the higher mussel -to-fish weight ratio than in tanks with lower 
ratios. This difference in weight of mussel per weight of fish is 
probably responsible for the difference in fish weight change between 
the two treatments. Again in trial eight temperature did not fluctuate 
as widely as in the previous trial; all three treatments lost less 
weight or gained more weight than the same treatments in trial seven. 
As in trial seven, fish held with thin-shell mussels in trial eight 
lost more weight than fish with thick-shell mussels perhaps in 
response to the increased weight ratio of mussel/fish in the thick- 
shell mussel treatment tanks. 



102 



The general lack of significant results in the feeding study is 
probably due to several • factors. First, fish probably do not benefit 
from the consumption of mussel feces under conditions normally found 
in the wild. Fish may consume this material but significant 
nutritional benefit from the forage was not indicated by this study. 
Although the water was strained through fine netting there were 
zooplankton and occasional insect nymphs and larvae present in the 
tanks. The chance appearance and consumption of these forage items in 
a tank could decrease the weight loss of the fish in that tank in 
comparison to tanks in which the fish did not obtain extra forage. 
Second, the slight weight changes among treatments may have been too 
small to register a significant difference with the sample sizes we 
were using. The results of the feces feeding trials suggest that 
mussel feces and pseudofeces are not a valuable forage for fathead 
minnows. 



103 



SUMMARY AND CONCLUSIONS 

This study reviewed the literature on interactions between fish 

and mussels and used three approaches (substrate preference tests, 

predator-prey trials, and feeding trials) in a laboratory setting to 

examine whether one forage fish (fathead minnow) and five common sport 

fishes (channel catfish, largemouth bass, bluegill, white crappie, and 

sauger) were more attracted to mussel beds than to other types of 

substrate, and if they were, to examine the basis for that attraction. 

The design and results of the laboratory experiments are summarized 

below. We attempted to verify the laboratory results in the field, 

using: (1) substrate preference tests conducted in raceways in Quiver 

Creek, near Havana, Illinois, and (2) fish population sampling in areas 

with and without mussel beds in the Mississippi River 50 miles upstream 

from St. Louis, Missouri. Repeated floods interrupted the raceway 

tests, which are not discussed further. Results of the fish population 

sampling were inconclusive because of sampling limitations (see Appendix 

G), but an improved sampling design is included in the recommendations 

at the end of this section. We fortuitously observed two phenomena that 

were incidental to the objectives of this research and that consequently 

are described in appendices: (1) the mantle flap display in the yellow 

sand shell, Lampsilis teres . (Appendix A)--the first (to our knowledge) 

videotaping and description of the way this mussel attracts its fish 

hosts, and (2) the remarkable survival of freezing by seven species of 

mussels (Appendix D) . Information from the literature search that 

should interest other researchers and managers is summarized in three 

tables: Table 1 lists the species of fish that reportedly associate 

with mussel beds, together with their food habits and status as 

104 



glochidial hosts; Table 2 lists invertebrates found in association with 
mussels; and Appendix C the filtration rates of some marine and 
freshwater bivalves, including the introduced Asiatic clam ( Corbicula 
fluminea ) and zebra mussel ( Dreissena polvmorpha ). 

Laboratory Studies 

Substrate preference trials determined whether the fish spent more 
time over mussels than over other types of substrate (cobbles and sand) 
and tested two hypotheses: (1) fish are attracted to the algae and 
small invertebrates attached to the mussel shells (the forage hypothe- 
sis), and (2) fish are attracted by the physical structure provided by 
the mussel shells (the structure hypothesis). The mussels were 
partially buried in sand in a natural position, with their pseudosiphons 
and the posterior portions of their shells above the sand surface. The 
cobbles were approximately the same size as the mussels and similarly 
placed in sand. Individual sport fish and individuals and schools of 
fathead minnows were given a choice of two substrates at a time in 
laboratory tanks. 

The forage hypothesis was tested by offering a choice between 
substrates colonized with attached algae and invertebrates and 
uncolonized substrates without attached organisms. The colonized 
substrates were cobbles and mussels from streams. Uncolonized 
substrates were mussels whose shells were scrubbed clean and cobbles 
obtained from roadsides or a quarry. The structure hypothesis was 



105 



tested by giving fish choices between bare sand and uncolonized cobbles 
or mussels. To check whether the shells alone were the attraction, 
rather than the living mussels, three species of fish (bluegill, white 
crappie, and juvenile channel catfish) were given a choice of empty 
mussel shells, as well as cobbles and live mussels. Empty shells were 
stored dry until placed in the preference tanks, and consequently 
there were no living organisms attached to them. 

The predator-prey trials measured the rate at which sauger 
captured fathead minnows over bare sand versus uncolonized cobbles. In 
each of six trials five fathead minnows were introduced on two 
consecutive days to a substrate preference tank containing a single 
sauger. 

The feeding studies stemmed from an observation during the 
preference trials of fathead minnows consuming mussel feces and a second 
observation of fish in the holding tanks eagerly consuming packets of 
glochidial larvae released by mussels. In the feeding trials, fish were 
held in aquaria alone or with mussels. The only food source provided 
during the eight two-week trials was from creek water continuously 
pumped into the aquaria. In every trial but one, the creek water was 
strained through a 0.01-inch mesh screen to remove most vegetation and 
invertebrates that were large enough for the fish to consume directly. 
The screens did not remove small particles that mussels are capable of 
filtering from the water. Thus the feces from the mussels were the 
major source of food, and the utilization and nutritional value could be 
judged by comparing the weight change of fish held with mussels to fish 



106 



held alone. In addition, the protein content of glochidial packets was 
measured to determine their potential nutritional value to fish. 

Substrate preference. The five species of fishes preferred 
uncolonized, live mussels in 17 of 22 tests and uncolonized cobbles in 
only four of 22 tests. Because of time constraints, only one preference 
test for empty, uncolonized mussel shells was run with each of three 
fish species. Bluegill and channel catfish preferred empty mussel 
shells while white crappie avoided them; only the avoidance was 
statistically significant at P < 0.05. The greater preference for 
live mussels over rocks and empty mussel shells indicates that the 
fish are not merely selecting bottom substrate because of the 
structure it provides. The preference also cannot be explained by the 
presence of small invertebrates or algae on the mussel shells, which 
had been scrubbed. 

Forage preference. Fish preferred colonized mussels in 19 of 22 
tests (86%); colonized cobbles seven out of ten times (70%); and, as 
mentioned above, uncolonized cobbles only four out of 22 times (18%). 
The greater preference for colonized cobbles in comparison to uncolo- 
nized cobbles supports the forage hypothesis, as does our observation 
that the fish eagerly picked at the invertebrates and filamentous algae 
on the colonized substrates. However, there must be some attraction 
associated with the live mussels themselves rather than the forage 
attached to them, because there were no statistically significant 
differences (P > 0.05) in preference among mussels with cleaned shells 
and mussels or cobbles with attached algae and invertebrates. Thus both 



107 



the substrate and forage tests consistently indicate some attractant 
associated with live mussels; or conversely, some repellent feature of 
uncolonized cobbles and uncolonized, empty shells. A common feature of 
the latter two substrates, besides having no attached live 
macroinvertebrates or filamentous algae, is that they were stored dry. 
All the other substrates were obtained from rivers and kept in water, so 
that their attached colonists remained alive. Some living algae and 
bacteria could have remained in the crevices near the umbones and hinge 
of even the scrubbed mussels. Marzolf (1966) noted that the amphipod 
Pontoporeia affinis preferred substrates whose surfaces had been 
"conditioned" by accumulated organic matter or bacteria, and it is 
possible that fish respond the same way. The possibility that mussels 
or their colonists release odors that attract fish was not investigated, 
but should be the subject of additional research. Although at least two 
of the 14 mussel species we used are known to employ visual lures 
(modified mantle flaps or glochidial packets that mimic minnows, worms, 
or other forage) to attract their fish hosts, none of the individuals 
were observed displaying during the preference trials. 

Predator-prey trials. The minnows were twelve times more 
vulnerable to sauger when the minnows were over bare substrate than 
when over cobbles, and consequently they spent twice as much time over 
the cobbles. Sauger spent an average of 3.8 minutes per minnow 
capture over sand but required 11 minutes to capture a minnow over the 
cobble. The minnows presumably would use mussels similarly as a 
refuge from predation, although time did not permit testing that 
hypothesis in additional predation trials. 



108 



Feces as forage. Fathead minnows did not appear to benefit from 
the consumption of mussel feces and pseudofeces: in five of the eight 
feeding trials fish held alone exhibited less weight loss or greater 
weight gain than fish held with mussels. 

Nutritional value of qlochidial packets. Though mussel feces may 
not be nutritious forage for fish, the glochidial packets are relatively 
high in protein. The protein content of aborted glochidial packets from 
two mapleleaf ( Quadrula quadrula ) and two pocketbook mussels ( Lampsil is 
cardium ) ranged from 55% to 77% on a dry weight basis and the embryonic 
larvae in the packet had scarcely any shell material. The protein 
content dropped to 20% in mature packets from a yellow sandshell mussel 
( Lampsilis cardium ), because glochidial shell material comprised a 
greater portion of the total mass. 

Importance of Mussel Beds to Fishes 

This laboratory study demonstrated that five common sport fishes 
(channel catfish, largemouth bass, bluegill, white crappie, and sauger) 
and one forage fish (fathead minnow) were more attracted to mussel beds 
than to other types of substrate. The results suggest that mussels 
serve as both direct and indirect sources of forage for game fish. The 
glochidial packets released by the mussels could provide a seasonally 
abundant food reward for the fishes that disperse the larval mussels. 
Young game fish probably consume the invertebrates that colonize mussel 



109 



shells, or the link may be from invertebrates to small fish, such as 
fathead minnows, to the piscivores. Sephton et al . (1980) found higher 
densities of invertebrates associated with mussel beds than with other 
substrates (the analysis included worms, insects, and snails, but 
excluded the mussels themselves). Another reason that small fish may 
concentrate in mussel beds, besides the presence of small invertebrates, 
is to avoid predators, as the fathead minnows did in our tests. 

Small invertebrates may be more abundant in mussel beds than in 
other substrates because: (1) the mussels increase the surface 
roughness of the bottom, thereby creating vertical eddies that bring 
food particles in the water column into the feeding range of bottom- 
dwelling organisms (Holloway 1990); (2) feces and pseudofeces of the 
mussels provide nutritious forage (Izvekova and Lvova-Katchanova 1972); 
(3) the shells are a solid attachment site for eggs, pupae, and feeding 
nets (Anderson and Vinikour 1984); (4) the interstices among the shells 
provide a refuge from predators and water currents and a collecting 
place for food particles, including mussel feces; and (5) the movements 
of the mussels slowly mix the top foot or so of the substrate, probably 
aerating the sediment and influencing exchanges of nutrients and organic 
matter between sediments and the water column. 

From the fishes' point of view, mussels serve another very 
important purpose besides providing foraging sites and refuge. 
Lithophilic spawners, such as walleye, apparently use both mussels and 
cobbles as spawning substrates (Balon 1975; Pitlo). Although none of 
our North American fishes are known to use mussels as brood chambers for 



110 



their eggs, several species elsewhere do, including at least 13 cyprinid 
species, a family with many North American representatives (Balon 1975). 

Mussels can serve as solid substrate and refuges for other 
invertebrates and fish because most native mussels continually expose a 
portion of their shells above the sediment surface, rather than 
completely burying themselves. Although cobbles can perform the same 
functions, cobbles can be covered by silt or sand whereas the living 
mussel actively maintains its position at the sediment surface. The 
supply of solid substrate continually accumulates through the death of 
individual mussels. This is a biogenic habitat (sensu Woodin 1978), 
capable of renewing and maintaining itself despite some environmental 
perturbation. The solid structure provided by mussel beds is likely to 
be most critical as a substrate in alluvial rivers otherwise dominated 
by shifting deposits of sand or mud. 

Recommendations 

1. Preservation and restoration of mussel beds should be regarded as an 
essential part of fisheries management, in view of the importance 
of mussel beds as a self-maintaining biogenic habitat that 
provides spawning sites, forage, and refuge for fish. Biogenic 
habitats generally cost much less to maintain in the long run than 
artificial substrates introduced by man. 



Ill 



The value of mussel beds to fisheries has much to do with the 
behavior and shell morphology of the native mussel species, 
therefore preservation and recovery of native mussels should be 
part of fisheries management. Research is needed to assess the 
impacts of introduced species on native mussels and to develop 
strategies to minimize damage and prevent other invasions. 
Introduced species, such as the zebra mussel, may displace the 
native mussels without replacing their services to other organ- 
isms, including fish. If the zebra mussel overgrows native 
mussels, it will interfere with their feeding and their display 
and release of glochidial packets and change the size and shape 
of the interstices available to other invertebrates and small 
fishes; and it may filter sufficient sperm from the water to 
reduce the fertilization rate of native mussels. 

Reviews of permits for discharges or developments and assessments 
of damages following spills should include effects on mussels, 
which may be less resilient to stress than fish. Many native 
species of mussels require 5-12 years to reach sexual maturity 
and recruitment is very sporadic, so they usually are the last 
group of aquatic organisms to recolonize an area where their 
populations have been reduced or eliminated. Adult mussels 
cannot avoid toxicants or other stresses as readily as more 
mobile organisms can. 



112 



4. We recommend two approaches (a-b below) using field data to define 
the relationship between fish populations and mussel beds. 
Previous field studies of associations between fish and mussel 
beds suffer from the weakness that fish collections were not 
made over areas without beds, because the purpose of the 
collections was only to determine what fishes were serving as 
hosts for mussels. Those studies therefore are not appropriate 
for determining whether fish preferentially congregate over 
mussel beds, although our laboratory study and others provide 
several reasons why they might do so. Fish use of mussel beds 
may be highly seasonal, so field sampling should include the 
spring spawning season for fishes such as walleye that are 
lithophilic spawners, and a period including spring through 
summer when various species of mussels are attracting fish hosts 
and releasing glochidia. Fish sampling should target, or at 
least include fish hosts of mussels known to occur in the beds. 

(a) Correlation analysis using existing data sets. The association 
between fish populations and mussel beds should be investigated using 
existing fish data sets and correlation or regression analyses. The 
Illinois Department of Conservation has a long-term data set on fish 
populations. The proximity of these stations to known mussel beds will 
need to be quantified, and it may be necessary to update mussel sur- 
veys in some areas. 



113 



(b) Special surveys and field experiments. Quantitative sampling of 
both fish and mussels should be conducted in areas where there are 
mussel beds and other areas where the water quality and habitat are 
similar, but there are few mussels. A survey approach, with a large 
number of replicate samples, could be adopted, or an experimental 
approach, or both. Quantitative sampling techniques for streams and 
small rivers are fairly well developed; quantitative or semi- 
quantitative sampling of fish in large rivers could involve some more 
experimental techniques such as remote sensing using fixed arrays of 
sonar devices over areas with and without mussel beds, deepwater 
electrofishing, and benthic and midwater trawling. Construction 
projects where entire mussel beds have been transplanted to another 
area would be ideal experiments, and a mussel -fish interaction study 
of this type should be written into these permits. Although the 
disturbance associated with construction is a confounding factor at 
the mussel removal site, the addition of mussels to the otherwise 
undisturbed location provides an ideal field experiment, where before- 
and after- measurements can be planned. Lethal episodes (spills of 
nonpersistent toxicants, low dissolved oxygen levels, excessive 
temperatures, are to be avoided at all costs, but when they do occur 
such episodes provide an opportunity to evaluate the effects of 
mussels on fish populations. The ideal situation would be to have 
sampled mussels and fish fortuitously before the episode, but even 
without predi sturbance data, it is possible to measure fish 
populations in association with natural recovery or a planned 
restocking of the mussel bed. 



114 



5. Bioenerqetic and population modeling of selected native mussel 
populations and zebra mussels could determine whether mussels 
contribute significantly to the food base and would also be 
useful in determining effects of competitors, such as zebra 
mussels, on the energy balance and ultimate survival of the 
native mussels. Mussel population models that include natural 
and harvest mortality would provide a rational, quantifiable 
basis for regulating mussel harvests, habitat disturbance, and 
discharges that affect growth and recruitment. The bioenergetic 
and population models would also be helpful in determining where 
mussel relocations or restorations might best succeed. 



115 



LITERATURE CITED 

Anderson, R.V. and W.S. Vinikour. 1984. Use of molluscs as pupation 
sites by Oecetis inconspicua (Trichoptera: Leptoceridae) . 
Journal of Freshwater Ecology 2:417-422. 

Arey, L.B. 1932. The nutrition of glochidia during metamorphosis. 
Journal of Morphology 53:201-221. 

Balon, E.K. 1975. Reproductive guilds of fishes: a proposal and 
definition. Journal of the Fisheries Research Board of Canada 
32:821-864. 

Bardach, J.F., J.H. Ryther and W.O. McLarney. 1972. Aquaculture. The 
farming and husbandry of freshwater and marine organisms. Wiley- 
Interscience, John Wiley and Sons. New York. 

Beedham, G.E. 1970. A further example of an association between a 
chironomid (Diptera) larva and a bivalve mollusc. Entomologists 
Monthly Magazine 105:3-5. 

Bennett, D.H. and J.W. Gibbons. 1972. Food of largemouth bass 
( Micropterus salmoides ) from a South Carolina reservoir 
receiving heated effluent. Transactions of the American 
Fisheries Society 101:650-654. 



116 



Bodensteiner, L.R., W.M. Lewis, and R.J. Sheehan. 1990. Differences in 
the physical environment of the upper Mississippi River as a 
factor in overwinter survival of fish. Pages 109-117 in 
Restoration of midwestern stream habitat symposium. Rivers and 
Stream Technical Committee, North-Central Division, American 
Fisheries Society. 

Brett, J.R. 1957. The eye. Chapter II, part 1. Pages 121-154 in M. 
Brown editor. The physiology of fishes. Volume II. Behavior. 
Academic Press Inc. New York. 

Brinkhurst, R.O. 1974. The benthos of lakes. St. Martin's Press. New 
York. 

Brown, P.B. and E.H. Robinson. 1989. Comparison of practical fish 
feeds containing 26 or 30 % protein. Progressive Fish Culturist 
51:149-151. 

Burky, A.J. 1983. Physiological ecology of freshwater bivalves. Pages 
281-327 in W.D. Russel -Hunter editor. The mollusca, vol 6, 
ecology. Academic Press, Inc. New York. 

Carver, C.E.A. and A.L. Mallet. 1990. Estimating the carrying capacity 
of a coastal inlet for mussel culture. Aquaculture 88:39-53. 



117 



Chamberlain, T.K. 1934. The glochidial conglutinates of the Arkansas 
fanshell , Cvproqenia aberti (Conrad). Woods Hole Biological 
Bulletin 66:55-61. 

Clark, H.W. and S. Stein. 1921. Glochidia in surface towings. The 
Nautilus 25:16-20. 

Clarke, A.H. 1981. The freshwater molluscs of Canada. National Museums 
of Canada. Ottawa, Canada. 

Coker, R.E., A.F. Shira, H.W. Clark, and A.D. Howard. 1921. Natural 
history and propagation of freshwater mussels. Bulletin of the 
U.S. Bureau of Fisheries 37(893) :71-181. 

Crowder, L.B. and W.E. Cooper. 1982. Habitat structural complexity and 
the interaction between bluegills and their prey. Ecology 
63:1802-1813. 

Cummins, K.W. and G.H. Lauff. 1969. The influence of substrate 
particle size on the microdistribution of stream macrobenthos. 
Hydrobiologia 34:145-181. 

Davenport, D. and M. Warmuth. 1965. Notes on the relationship between 
the freshwater mussel Anodonta impl icata (Say) and the alewife 
Alosa pseudoharengus (Wilson). Limnology and Oceanography 10:74- 
78. 



118 



De Bruin, J.P.C. and C. Davids. 1970. Observations on the rate of 
water pumping of the freshwater mussel Anodonta cygnea zellensis 
(Gmelin). Netherlands Journal of Zoology 20:380-391. 

DeMarch, B.G.E. 1976. Spatial and temporal patterns in macrobenthic 
stream diversity. Journal of the Fisheries Research Board of 
Canada 33:1261-1270. 

Driscoll, E.G. and D.E. Brandon. 1973. Mollusc - sediment 
relationships in northwestern Buzzards Bay, Massachusetts, 
U.S.A. Malacologia 12:13-46. 

Eckblad, J.W. 1986. The ecology of pools 11-13 of the upper 
Mississippi River: a community profile. U.S. Fish and Wildlife 
Service Biological Report 85 (7.8). 

Egglishaw, H.J. 1964. The distributional relationship between the 
bottom fauna and plant detritus in streams. Journal of Animal 
Ecology 33:463-476. 

Ellis, M.M. 1929. The artificial propagation of freshwater mussels. 
Transactions of the American Fisheries Society 59:217-223. 



, 1931. Some factors affecting the replacement of the commercial 
fresh-water mussels. U.S. Bureau of Fisheries. Fishery Circular 
No. 7. 



119 



Eriksen, C.H. 1966. Benthic invertebrates and some substrate-current- 
oxygen interrelationships. Pages 98-115 in K.W. Cummins, C.A. 
Tryon and R.T. Hartman editors. Organism-substrate relationships 
in streams. Special Publication Number 4, Pymatuning Laboratory 
of Ecology, University of Pittsburgh. 

Forbes, S.A. and R.E. Richardson. 1908. The fishes of Illinois. 
Illinois Printing Co., Danville, IL. 



and . 1920. The fishes of Illinois. Illinois State Journal 

Co., State Printers. Springfield, IL. 



Fraser, D.F. and R.D. Cerri . 1982. Experimental evaluation of 
predator-prey relationships in a patchy environment: 
consequences for habitat use patterns in minnows. Ecology 
63:307-313. 

Fuller, S.L.H. 1.974. Clams and mussels. Pages 215-273 in C.W. Hart and 
S.L.H. Fuller editors. Pollution ecology of freshwater 
invertebrates. Academic Press. New York. 



. 1978. Freshwater mussels (Mollusca: Bivalvia: Unionidae) of the 
upper Mississippi River: observations at selected sites within 
the 9-foot channel navigation project on behalf of the United 
States Army Corps of Engineers. American Midlands Naturalist 
111:16-22. 



120 



and D.J. Bereza. 1974. The value of anatomical characters in 
naiad taxonomy (Bivalvia: Unionaceae). Bulletin of the American 
Malacological Union, Inc: 21-22. 



Gerking, S.D. 1962. Production and food utilization in a population of 
bluegill sunfish. Ecological Monographs 32:31-78. 

Gorman, O.T. 1988. An experimental study of habitat use in an 
assemblage of Ozark minnows. Ecology 69:1239-1250. 

Greier, N.M. 1926. Notes on the naiads of the upper Mississippi River, 
III. On the relation of temperature to the rhythmical 
contractions of the "mantle flap" in Lampsil is ventricosa 
(Barnes). Nautilus 39:111-114. 

Habel , M.L. 1970. Oxygen consumption, temperature tolerance and 

filtration rate of the introduced Asiatic clam Corbicul a 

manilensis from the Tennessee River. Masters Thesis, Auburn 
University. 

Hargeby, A. 1990. Macrophyte associated invertebrates and the effect 
of habitat permanence. Oikos 57:338-346. 

Hargrave, B.T. 1976. The central role of invertebrate feces in 
sediment decomposition. Pages 301-321 in J.M. Anderson and A. 
Macfayden editors. The role of terrestrial and aquatic organisms 
in decomposition processes. Blackwell, Oxford. 



121 



Harman, W.N. 1970. New distributional records and ecological notes on 
central New York Unionacea. American Midlands Naturalist 84:46- 
58. 

Haven, D.S. and R. Morales-Alamos. 1966. Aspects of biodeposition by 
oysters and other invertebrate filter feeders. Limnology and 
Oceanography 11:487-498. 

Holland-Bartels, L.E. 1990. Physical factors and their influence on 
the mussel fauna of a main channel border habitat of the upper 
Mississippi River. Journal of the North American Benthological 
Society 9:327-335. 

Holloway, M. 1990. Mussels may clump together to improve their food 
supply. Scientific American 263(6) :30. 

Hornbach, D.J., CM. Way, T.E. Wissing and A.J. Burky. 1984. Effects 
of particle concentration and season on the filtration rates of 
the freshwater clam Sphaerium striatinum Lemark (Bivalvia: 
Pisidiidae). Hydrobiologia 108:83-96. 

Howard, A.D. 1913. Experiments in propagation of freshwater mussels of 
the Quadrula group. U.S. Bureau of Fisheries Document No. 801. 

• 1914. Some cases of narrowly restricted parasitism among 

commercial species of freshwater mussels. Transactions of the 
American Fisheries Society 44:40-44. 

122 



_. 1917. An artificial infection with glochidia on the river 
herring. Transactions of the American Fisheries Society 46:93- 
100. 

_. 1922. Experiments in the culture of fresh-water mussels. 
Bulletin of the U.S. Bureau of Fisheries 38:63-89. 

. and B.J. Anson. 1922. Phases in the parasitism of the 

Unionidae. Journal of Parasitology 9:68-82. 

Hynes, H.B.N. 1970. The ecology of running waters. Liverpool 
University Press. Liverpool. 

Isley, F.B. 1911. Preliminary note on the ecology of the early 
juvenile life of the Unionidae. Biological Bulletin 20:77-80. 

Izvekova, E.I. and A. A. Lvova-Katchanova. 1972. Sedimentation of 
suspended matter by Drei ssena polvmorpha and its subsequent 
utilization by chironomid larvae. Polish Archives of 
Hydrobiology 19:203-210. 

Keeney, D.R. 1973. The ammonia cycle in sediment water systems. 
Journal of Environmental Quality 2:15-29. 



123 



Klar, G.T. and N.C. Parker. 1989. Evaluation of five commercially 
prepared diets for striped bass. Progressive Fish Culturist 
51:115-120. 

Kraemer, L.R. 1970. The mantle flap in three species of Lampsilis 
(Pelecypoda: Unionidae). Malacologia 10:225-282. 

Kranz, P.M. 1974. The anastrophic burial of bivalves and its 
paleoecological significance. Journal of Geology 82:237-265. 

Kryger, J. and H.U. Riisgard. 1988. Filtration rate capacities in six 
species of European freshwater bivalves. Oecologia 77:34-38. 

Lauritson, D.K. 1986. Filter-feeding in Corbicula fluminea and its 
effect on seston removal. Journal of the North American 
Benthological Society 5:165-172. 

LeFevre, G. and W.C. Curtis. 1912. Studies on the reproduction and 
propagation of freshwater mussels. Bulletin of the U.S. Bureau 
of Fisheries Document 756. 

Leff, L.G., J.L. Burch and J.V. McArthur. 1990. Spatial distribution, 
seston removal, and potential competitive interactions of the 
bivalves Corbicula fluminea and Elliptio complanata , in a plain 
stream. Freshwater Biology 24:409-416. 



124 



Lund, E.J. 1957. A quantitative study of clearance of a turbid medium 
and feeding by the oyster. Publication of the Institute of 
Marine Science, Texas 4:296-312. 

Marzolf, G.R. 1966. The trophic position of bacteria and their 
relation to the distribution of invertebrates. Pages 131-135 in 
K.W. Cummins, C.A. Tryon and R.T. Hartman editors. Organism- 
substrate relationships in streams. Special Publication No. 4, 
Pymatuning Laboratory of Ecology. University of Pittsburgh. 

Mathiak, H.A. 1979. A river study of the unionid mussels of Wisconsin 
1973 - 1977. Sand Shell Press. Horicon, Wisconsin. 

Maynard, L.A., J.K. Loosli, H.F. Hintz and R.G. Warner. 1979. Animal 
nutrition, 7th edition. McGraw-Hill Book Company. New York. 

McCullough, D.A., G.W. Minshall and C.E. Cushing. 1979. Bioenergetics 
of lotic filter-feeding insects Simul ium spp. (Diptera) and 
Hydropsyche occidental is (Trichoptera) and their function in 
controlling organic transport in streams. Ecology 60:585-596. 

Merritt, R.W. and K.W. Cummins. 1984. An introduction to the aquatic 
insects of North America, 2nd. edition. Kendall -Hunt Publishing 
Co. Dubuque, Iowa. 



125 



Miller, A.C., B.S. Payne and T. Siemsen. 1986. Description of the 
habitat of the endangered mussel Plethobasus cooperianus . The 
Nautilus 100:14-18. 

Mohlenberg, F. and H.U. Riisgard. 1979. Filtration rate, using a new 
indirect technique, in thirteen species of suspension-feeding 
bivalves. Marine Biology 54:143-147. 

Negus, C.L. 1966. A quantitative study of growth and production of 
unionid mussels in the Thames at Reading. Journal of Animal 
Ecology 35:513-532. 

Ney, J.J. 1978. A synoptic review of yellow perch and walleye biology. 
Pages 1-12 in R.L. Kendall editor. Selected coolwater fishes of 
North America. American Fisheries Society Special Publication 
11. American Fisheries Society. Washington, D.C. 

Ortmann, A.E. 1911. Monograph of the najads of Pennsylvania, parts I 
and II. Memoirs of the Carnagie museum 4:279-347. 

Paterson, C.G. 1984. A technique for determining apparent selective 
filtration in the freshwater bivalve Elliptio complanata 
(Lightfoot). The Veliger 27:238-241. 

• and I.F. Cameron. 1985. Comparative energetics of two 

populations of the unionid, Anodonta cataracta . Freshwater 
Invertebrate Biology 4:79-90. 



126 



Petridis, D. 1990. Utilization of grass carp feces by the isopod 
Asellus aquaticus (L.) in the laboratory. Hydrobiologia 194:47- 
56. 

Piper, R.G., I.B. McElwain, L.E. Orme, J. P. McCraren, L.G. Fowler, and 
J.R. Leonard. 1986. Fish hatchery management. U.S. Dept. of the 
Interior. Fish and Wildlife Service. Washington, D.C. 

Pitlo, J. Jr. 1989. Walleye spawning habitat in pool 13 of the upper 
Mississippi River. North American Journal of Fisheries 
Management 9:303-308. 

Rankin, E.T. 1986. Habitat selection by smallmouth bass in response to 
physical characteristics in a natural stream. Transactions of 
the American Fisheries Society 115:322-334. 

Reeders, H.H. and A. Bij de Vaate. 1990. Zebra mussels ( Dreissena 
polymorpha ) : a new perspective for water quality management. 
Hydrobiologia 200/201:437-450. 

Rhoads, D.C. and D.K. Young. 1970. The influence of deposit - feeding 
organisms on sediment stability and community trophic structure. 
Journal of Marine Research 28:150-178. 



127 



Robinette, H.R. 1976. Effect of selected sublethal levels of ammonia 
on the growth of channel catfish ( Ictal urus punctatus ) . 
Progressive Fish Culturist 38:26-29. 

SAS Institute Inc. 1985. SAS/STAT guide for personal computers, 
Version 6 Edition. SAS Institute Inc. Cary, North Carolina. 

Scott, W.B. and E.J. Crossman. 1973. Freshwater fishes of Canada. 
Bulletin 184. Fisheries Research Board of Canada, Ottawa. 

Sechnick, C.W., R.F. Carline, R.A. Stein and E.T. Rankin. 1986. 
Habitat selection by smallmouth bass in response to physical 
characteristics of a simulated stream. Transactions of the 
American Fisheries Society 115:314-321. 

Segler, W.F. 1963. Fishes of Utah. Utah Dept. of Fish and Game. Salt 
Lake City. 

Sephton, T.W., C.G. Paterson, and C.H. Fernando. 1980. Spatial 
interrelationships of bivalves and nonbivalve benthos in a small 
reservoir in New Brunswick, Canada. Canadian Journal of Zoology 
58:852-859. 

Sheehan, R.J., W.M. Lewis and L.R. Bodensteiner. 1990. Winter habitat 
requirements and overwintering of riverine fishes. Dingell- 
Johnson Wallop-Breaux Sport Fish Restoration Project Completion 
Report F-79-R. Carbondale, Illinois. 

128 



Shpigel, M. and R. Fridman. 1990. Propagation of the manila clam 
( Tapes semidecussatus ) in the effluent of fish aquaculture ponds 
in Eilat, Israel. Aquaculture 90:113-122. 

Smith, P.K., R.I. Krohn, G.T. Hermanson, A.K. Mall ia, F.H. Gartner, 
M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson and D.C. 
Klenk. 1985. Measurement of protein using bicinchoninic acid. 
Analytical Biochemistry 150:76-85. 

Smith, P.W. 1979. The fishes of Illinois. University of Illinois 
Press. Chicago. 

Starrett, W.C. 1950. Food relationships of the minnows of the Des 
Moines River, Iowa. Ecology 31:216-233. 

. 1971. A survey of the mussels (Unionacea) of the Illinois 

River: a polluted stream. Illinois Natural History Bulletin 
30:266-403. 

Stein, R.A. 1977. Selective predation, optimal foraging, and the 
predator-prey interaction between fish and crayfish. Ecology 
58:1237-1253. 



. 1979. Behavioral response of prey to fish predators, in R.H. 
Stroud and H. Clepper editors. Predator-prey systems in 
fisheries management. Sport Fishing Institute, Washington, D.C. 



129 



and J.J. Magnuson. 1976. Behavioral response of crayfish to a 
predator. Ecology 57:751-761. 



Stern, E.M. 1978. Identification of host fishes for four species of 
freshwater mussels (Bivalvia: Unionidae). American Midlands 
Naturalist 100:233-236. 

Stickney, R.R. 1979. Principles of warmwater aquaculture. John Wiley 
and Sons, Inc. New York. 

Storey, K.B. and J.M. Storey. 1990. Frozen and alive. Scientific 
American, December: 92-97. 

Taylor, B.R. and J.C. Roff. 1984. Use of ATP and carbon:nitrogen ratio 
as indicators of food quality of stream detritus. Freshwater 
Biology 14:195-201. 

Tedla, S. and C.H. Fernando. 1969. Observations on the glochidia of 
Lampsil is radiata (Gmelin) infesting yellow perch, Perca 
flavescens (Mitchell) in the Bay of Quinte, Lake Ontario. 
Canadian Journal of Zoology 47:705-712. 

Vinyard, G.L. and W.J. O'Brien. 1976. Effects of light and turbidity 
on the reactive distance of bluegill ( Lepomis macrochirus ) . 
Journal of the Fisheries Research Board of Canada 33:2845-2849. 



130 



Waller, D.L., L.E. Holland, L.G. Mitchell and T.W. Kammer. 1985. 
Artificial infestation of largemouth bass and walleye with 
glochidia of Lampsil i s ventricosa (Pelicypoda: Unionidae) 
Freshwater Invertebrate Biology 4:152-153. 

Welsh, J.H. 1969. Mussels on the move. Natural History 78(5) :56-59. 

Wetzel, R.G. 1983. Limnology 2nd edition. Saunders College Publishing. 
Chicago. 

Widdows, J., P.Fieth and CM. Worrall . 1979. Relationships between 
seston, available food and feeding activity in the common mussel 
Mvtilus edulis . Marine Biology 50:195-207. 

Wickler, W. 1978. Mimicry in plants and animals. McGraw-Hill. New 
York. 

Wilbur, R. L. 1969. The redear sunfish in Florida. Fishery Bulletin 
No. 5. Florida Game and Freshwater Fish Commission. Dingall - 
Johnson Project F-22. 

Wiles, M. 1975. The glochidia of certain Unionidae (Mollusca) in Nova 
Scotia and their fish hosts. Canadian Journal of Zoology 53:33- 
41. 



131 



Wilson, C.B. and H.W. Clark. 1912. The mussel fauna of the Kankakee 
basin. U.S. Bureau of Fisheries Document 758:1-52. 

Winter, J.E. 1978. A review on the knowledge of suspension feeding 
lamelli branchiate bivalves with special reference to artificial 
aquaculture systems. Aquaculture 13:1-33. 

Wisniewski, R. 1990. Shoals of Dreissena polymorpha as bio-processors 
of seston. Hydrobiologia 200/201:451-458. 

Woodin, S.A. 1978. Refuges, disturbance, and community structure: a 
marine soft bottom example. Ecology 59:274-284. 

Zale, A.V. and R.J. Neves. 1982. Fish hosts of four species of 
lampsiline mussels (Mollusca: Unionidae) in Big Moccasin Creek, 
Virginia. Canadian Journal of Zoology 60:2535-2542. 



132 



APPENDICES 



133 



Appendix A. Yellow sand shell mantle flap display 

On 31 March 1990 we collected about two dozen yellow sand shell 
mussels ( Lampsilis teres ) from a mudflat in pool 26 of the Mississippi 
river near Alton, Illinois. The mudflat had become exposed when the 
pool was dewatered for construction of the new lock and dam. We 
collected the mussels for use in the preference study and held them in 
outdoor tanks. 

In June 1990 the yellow sand shells began to display their 
mantle flaps. We first noticed the display on the afternoon of 26 
June. The observations in the present study are drawn from three hours 
of video tape recorded 27 - 28 June and 5 - 6 July 1990 of two mussels 
filmed in the outdoor holding tanks. 

Fuller and Bereza (1974) described the structure of the yellow 
sand shell mantle flap from preserved specimens and compared it to L 
fasciola . In L^. teres there is no anterior eyespot or posterior tail 
differentiation as there is in L± fasciola . L^ cardium and other 
members of the genus. The two individuals filmed showed some variation 
in the margin of the mantle flap; one had a smooth mantle margin while 
the other was fimbriated. Fuller and Bereza do not mention this 
variation in the mantle margin, but do suggest that mantle characters 
may be a means for distinguishing between very similar species. 



134 



The flapping behavior of U_ teres has not been previously 
described but in her thorough study of lampsilid flapping behavior 
Kraemer (1970) discussed the mantle structure and behavior of L_s. 
cardium , l_^ sil iquoidea and l^ reeviana brevicula . Her terminology is 
used to describe the flapping behavior and positioning movements of 
the mussels. 

Like cardium the sand shells performed "preparatory " movements 
at the beginning of the flapping display. Over a period of about 40 to 
50 minutes the mussel moves from the normal position and assumes the 
headstand position (Figure 32), with the foot spread out on the 
surface of the sediment to prop the mussel up. Flap movements begin 
weakly and irregularly at first, then become stronger and occur more 
regularly and at shorter intervals. In L. teres both marsupia always 
protruded between the flaps, while L. cardium displayed one or both 
marsupia (Kraemer 1970). 

Kraemer categorized the flapping behavior of L. cardium into 
regular and slow flapping movements. Regular movements have high 
flapping frequencies of 60 or more moves per minute, slow movements 
occur at rates of less than 30 moves per minute. The flap movements we 
observed in L. teres are very similar to the slow movements Kraemer 
describes for L. cardium. She states: 



135 



Rotation 



Water 



Excurrent 
siphon 



Incurrent 
siphon 




Incurrent 
siphon 

Excurrent 
siphon 



Water 




Marsupium 
containing mussel larvae 



Mantle flap 



Foot 



Sediment 



Figure 32. Repositioning movement of a yellow sandshell mussel to assume 
the headstand position for the mantle flap display. 



136 



"Before the slow movement starts the flaps are spread 
wide apart, the entire length floating out horizontally, 
inner sides uppermost in the water. The marsupium may not 
but more often does, protrude between the flaps. When the 
movement begins, there is a contraction at the flap base; 
the tails move up and touch medially; then a pulse moves 
from in front of the tails, which draws the eyespot ends of 
the flap upright, together and backward" (Kraemer 1970 p. 
243). 



The basic inward movement of the flap described above is the same; 
the recovery movement is also similar. 

"In recovery, first the tails, then gradually the rest 
of the flaps relax and float out horizontally once more. At 
the end of the recovery stroke, the flaps have moved forward 
slightly again"' (Kraemer 1970 p. 243). 

The display in L* teres differs from L. cardium in that during the 
height of the display, the flaps are firmly pressed against the sides of 
the marsupia and alternately contracted anteriorly then posteriorly two 
or three times in unison so that both marsupia move with the flaps. 

The marsupia themselves are composed of numerous glochidial 
conglutinates and thus appear segmented. The marsupium has a white base 
color with gray pigmentation distally. One individual had a single 
brilliant white spot dorsally in the center of each marsupium. The 
shaking or quivering movement of the marsupium caused by the flaps 
further serves to give the impression of a moving worm or insect larva. 

The series of contractions - inward movement, antero-posterior 
contractions and relaxation of the flaps takes about two to three 
seconds; eleven to twelve series of movements occur in a minute. 



137 



A total of eleven yellow sand shells mussels displayed their 
mantle flaps over a two- to three-week period, though most of the 
mussels displayed only at night. This contrasts with the rapid flapping 
behavior of Lj. cardium observed by Kraemer. She reports that most of the 
flapping took place during the day, beginning at sunrise and ending at 
sunset, but the slow movements she observed took place at low light 
levels over long periods of time. 

The earliest the two most frequently observed mussels began 
flapping was 08:30 and other mussels which began later, after sunset, 
flapped at least as late as 23:30, though the flapping behavior always 
ceased by morning. The mussels ceased their displays during the second 
week of July. Morning water temperature in the tanks averaged 73 °F and 
dissolved oxygen averaged 6.7 ppm. Ammonia and pH were measured weekly, 
pH was 8.9, mean un-ionized ammonia concentrations was 0.16 ppm. 

Production of the glochidia and the flapping behavior may be 
energetically costly for the mussels. The marsupium of one individual 
comprised 44 percent of the total dry tissue weight of the mussel. The 
mussels which performed the display over the two- to three-week period 
died shortly after the display activities ceased. One possible reason 
for the mortalities may be that the mussels were unable to disperse the 
glochidia since fish were not present in the tanks. We placed two white 
crappies in the tank with the mussels, but the fish were more interested 
in nest construction than the mantle flap display. 



138 



In addition to Kraemer, other authors (Ortmann 1911, Wilson and 
Clark 1912, Coker et al . 1921, Howard and Anson 1922, Grier 1926, Welsh 
1969) have discussed mussel mantle flap behavior and theorized on its 
function. Two main hypotheses surface repeatedly: attraction of a fish 
host and aeration of the glochidia. We feel the minnow-like mantle flap 
of Lt cardium and the larvae- or worm-like appearance of the L. teres 
marsupia tend to support the host attraction hypothesis. 

This highly visual mantle flap display probably helps attract 
sight feeding hosts. Hosts for the yellow sand shell include white 
crappie, largemouth bass, smallmouth bass and other centrarchids. Turbid 
water can interfere with fish feeding (Vinyard and O'Brien 1976) and 
thus may interfere with the reproduction of lampsilid mussels but the 
flapping behavior could attract a potential host by creating pressure 
waves that fish could sense with their lateral line systems. 



139 



Appendix B. Protein content of glochidial conglutinates. 

Glochidial packets from three mussel species were collected from the 
outdoor holding tanks at the Forbes station. Four samples, two each from 
mapleleaf ( Quadrula auadrula ) and pocketbook ( Lampsilis cardium ) mussels 
were aborted packets shed after handling and transportation. The other 
two samples were mature packets from recently dead yellow sand shell 
mussels (L^ teres ). One of these mussels was collected as it was dying, 
the other was found after it had been dead for a day, withthe packets 
lying next to the mussel. 

The protein content of the six glochidial conglutinate samples was 
measured using two methods: the bicinchoninic acid (BCA) microprotein 
analysis (Smith et al . 1985) that we used to analyze sediment protein 
in the feeding trials, and Kjeldahl protein analysis. 

We extracted between 1.06xl0" 5 and 1.52xl0" 5 lb of conglutinate 
for each specimen for the BCA analysis and used two subsamples of the 
extracted sample for each individual. Since the quantity of the aborted 
packet samples was small, the mapleleaf and pocketbook packets were 
pooled by species for the Kjeldahl analysis. Sample size ranged from 
8.59xl0" 6 lb for the pocketbook mussels to 4.59xl0" 4 lb for the yellow 
sand shell mussels. The protein content measured by the BCA analysis 
ranged from 5.05 to 14.8 percent. Crude protein measured by Kjeldahl 
analysis ranged from 6.94 to 77.23 percent. Table 10 presents results of 



140 



Table 10. Percent protein content of glochidial conglutinates as 
determined by two methods. Means of BCA analysis are given with 
(subsample measurements) . 



Analysis 
Species Specimen Kjeldahl BCA 



L. teres 



L. cardium 



Q. quadrula 



21 
21 


.52 
.67 


12.50 a 1 
(12.50, 12.50) 


6 


.94 


5.435 c 
(5.05, 5.82) 


77 


,23 2 


11.055 ab 
(10.73, 11.38) 

14.655 a 
(14.48, 14.83) 


53 


.48 3 


8.595 be 
(8.44, 8.75) 

11.155 ab 
(9.56, 12.75) 



1) Means with different letters are significantly different (P < 0.05) 

2) Specimens three and four pooled for Kjeldahl analysis. 

3) Specimens five and six pooled for Kjeldahl analysis. 



141 



the two analyses. 

The Kjeldahl protein analysis indicated aborted packets were 
higher in percent protein than mature packets. This difference may be 
due to the presence of shell material in the advanced conglutinates, 
which adds mass but not protein. 

ANOVA indicated significant differences in percent protein 
between individuals but not between species. This suggests that there 
was no significant difference between mature and aborted packets, but 
that packets collected after sitting on the bottom of the tank for a 
day, (specimen two) had significantly lower protein content than fresh 
packets from the same species (specimen one). 

The total percent protein measured by the BCA analysis tended to 
be less than that measured by the Kjeldahl analysis. This difference is 
probably due to the relatively mild extraction technique we used with 
the BCA method versus that used in the Kjeldahl analysis. The Kjeldahl 
method uses sulfuric acid to extract nitrogen from the proteins (Maynard 
et al . 1979); for the BCA analysis we immersed the sample in one percent 
solution of sodium lauryl sulfate (SDS), a detergent, at 140 °F for four 
hours. 

It is reasonable to assume the percent crude protein values from 
the Kjeldahl analysis are absolute protein values, therefore by 
comparison, the BCA method extracted from 11 to 84 percent of the total 
protein present in the samples. This suggests that when used in this 
manner, one percent SDS is an inefficient extraction solution for 

142 



protein in glochidial conglutinates. 

The results do provide basic information on the protein content of 
glochidial packets and indicate that aborted conglutinates contain 
protein comparable to that in prepared fish diets (Klar and Parker 1989, 
Brown and Robinson 1989, Stickney 1979). Mature conglutinates, though 
lower in protein, could form a seasonally abundant supplemental food 
source for fish, but the nutritional value of the packet appears to 
diminish quickly when exposed to leaching effects of water. 



143 



Appendix C. Filtration rates of some marine and freshwater 
bivalves. 



Species 


Filtration 


rate/10" 4 


Reference 


Anodonta cygnea 


0.9-1.1 


gal/h/lb a 


DeBruin and Davids 1970 


A. cataracta 


0.8-5.0 


gal/1 b/h 


Paterson and Cameron 1985 


Elliptio complanata 
ii 

n 


28.8 

0.2 

1.9 


gal/muss/h 

gal/h/lb 

gal/h/lb 


Leff et al . 1990 
Paterson 1984 



Dreissena polymorpha 



Corbicula fluminea 
•I 

■I 

Sphaerium striatinum 

Mytilus edulis 
■I 

Cardium echinatum 
C. edule 

Modiolus modiolus 
Arctica islandica 



5.3-75.8 gal/clam/h 
0.2 gal/h/lb 

132-158 gal/clam/h 



114.0 gal/clam/h 

29.0 gal/clam/h 

3619 gal/clam/h 



Widdows et al . 1990 
Izvekova and Lvova- 

Katchanova 1972 
Reeders and 

Bij de Vaate 1990 

Leff et al . 1990 
Habel 1970 
Lauritson 1986 



0.15-22.1 gal/clam/h Hornbach et al . 1984 



6.98-14.5 gal/h/lb 

8.9-39.4 gal/h/lb. 

21.9 gal/h/lb b 

14.2 gal/h/lb 

32.3 gal/h/lb 
22.3 gal/h/lb 
16.8 gal/h/lb 



Carver and Mallet 1990 
Widdows et al . 1990 
Mohlenberg and Riisgard 1979 



a) Estimated using filtration rate reported by DeBruin and Davids with 
dry weight conversion factor for A. grandis from the present study. 

b) Rates calculated using equations in Mohlenberg and Riisgard based on 
0.002 pound body weight. 



144 



Appendix D. Survival of frozen mussels. 

The outdoor tanks where we held the mussels froze solid early in 
the winter of 1990 before we could move the mussels to Quiver Creek. 
Upon cleaning the tanks in the spring we were surprised to find that 
some of the mussels had survived. We are certain the tanks froze 
completely since they were on the surface of the ground, are not 
protected from the weather and not insulated in any way. 

The surviving mussels were not of a single species or size class. 
(Table 11). Two floaters, a three-ridge and a white heel - 
splitter were quite large, shell length > 4.9 inches; and there was one 
young floater and one young washboard mussel. The other survivors were 
mature individuals of various species averaging 2.3 inches in length. 

IntertidaT mussels survive freezing when exposed to air at low 
tide in northern latitudes (Storey and Storey 1990). Freezing is an 
unlikely event on a large river, though temperatures in the main channel 
are near freezing for much of the winter (Sheehan et al . 1990). Shallow 
backwater areas can freeze to the bottom in a severe winter 
(Bodensteiner et al . 1990). Since mussels are adapted to survive long 
periods of near freezing temperatures in main channel habitats. It is 
possible that they have evolved adaptations for survival of freezing. 



145 



Table 11. Number of mortalities and survivors of mussels frozen in 
outdoor tanks. 



Species 1 


Mortalities 


Survivors 


Percent 
Survival 


Amblema pi icata 


8 


2 


20.0 


Anodonta grandis 


34 


4 


10.5 


Arcidens confragosus 


1 


- 


0.0 


Lampsilis teres 


1 


1 


50.0 


Lasmigona complanata 


7 


1 


12.5 


Leptodea fragilis 


1 


- 


0.0 


Megalonaias nervosa 


3 


1 


25.0a 


Obliquaria reflexa 





2 


100.0 


Potamilus alatus 


1 


- 


0.0 


Quadrula quadrula 


20 


5 


20.0 


Q. pustulosa 


2 


- 


0.0 


Total 


78 


16 


17.02 



a) Two adult and one juvenile washboard died, the survivor was a 
juvenile. 



146 



Appendix E. Total time (fish-minutes) spent over sand and rock in 
the predator-prey trials by sauger and fathead minnows. 

Fish-minutes equal the number of fish present over a given 
substrate during a time interval multiplied by the length of the 
interval. Trials began with five minnows and one sauger. The adjustment 
factor adjusts the length of the trial to a mean duration by species for 
calculation of mean percent time spent over each substrate for each 
species. 



Sauger 














Adj 


usted 




Fish 


-minut 


es 


Mean 


Adjustment 


Fish- 


minutes 


Trial 


Sand 


Rock 


Total 


Total 


Factor 


Sand 


Rock 


1 


7 


8 


15 


35 


2.33 


16 


19 


2 


48 


14 


62 


35 


0.56 


27 


8 


3 


48 





48 


35 


0.73 


35 





4 


6 


23 


29 


35 


1.21 


7 


28 


5 


10 


1 


11 


35 


3.18 


32 


3 


6 


10 


36 


47 


35 


0.76 


8 


27 



Fathead Minnow 















Adj 


usted 




Fish 


-minutes 


Mean 


Adjustment 


Fish- 


minutes 


Trial 


Sand 


Rock 


Total 


Total 


Factor 


Sand 


Rock 


1 


17 


28 


45 


72 


1.60 


27 


45 


2 


130 


74 


204 


72 


0.35 


46 


26 


3 


38 


45 


83 


72 


0.87 


33 


39 


4 


3 


26 


29 


72 


2.49 


8 


64 


5 


4 


11 


15 


72 


4.81 


19 


53 


6 


8 


49 


57 


72 


1.27 


10 


62 



147 



Appendix F. Calculation of percent dry tissue weight from whole 
wet body weight for mussels used in the feeding trials. 

Mean whole and partial wet and dry weights for the sample of ten 
three ridge (thick) and ten floater (thin) mussels. Mean shell length: 
three-ridge 4.49 inches, floater 3.56 inches. Length measured as maximum 
antero-posterior. dimension of the shell. All weight measurements in 
pounds. 



Three ridge 



Floater 



Total Wet Weight 

Tissue Wet Weight 
Shell Wet Weight 
Water Weight 3 

Total Dry Weight 

Tissue Dry Weight 
Shell Dry Weight 
Water Weight 13 

Percent Dry Weight c 

Percent Dry Tissue^ 

Percent Dry Shell 

Percent Moisture 



0.79 



0.24 



0.12 
0.55 
0.13 


0.07 
0.04 
0.12 


0.55 


0.04 


0.016 
0.534 
0.24 


0.004 
0.038 
0.20 


69.62 


16.67 


2.02 (0.6515) 


1.66 (.1476) 


67.59 


15.83 


30.38 


83.33 



a) Water contained in mantle cavity. 

b) Difference between wet and dry total weights. 

c) Measured as (weight)/(total wet weight). 

d) Used as conversion factor for total wet weight to dry tissue weight. 
Standard deviation in parenthesis. 



148 



Appendix G. Methods and results of field sampling at pool 26 of 
the Mississippi River. 

We identified three sites at the upper end of pool 26 of the 
Mississippi River as non-mussel, mussel sanctuary and exploited mussel 
bed areas. The non-mussel site was located on the west bank at river 
mile 240, the sanctuary site was on the east bank at river mile 238.4 
and the exploited site was on the east bank at river mile 233.5. The 
sites were sampled from 20 through 23 August 1991. 

Substrate at each site was sampled with a petite Ponar dredge and 
the presence or absence of mussels was confirmed using a five foot brail 
bar. Current velocity, temperature, dissolved oxygen, pH, conductivity 
and alkalinity were sampled at the surface daily at each site. Water 
clarity was measured with a Secchi disk. 

Physico-chemical Parameters. Mean alkalinity, pH, conductivity and 
current velocity did not differ significantly between the three sites. 
Mean Secchi disc readings were significantly greater at the non-mussel 
site than at the sanctuary site. Mean Secchi disc reading at the 
exploited bed did not differ from either of the two other sites. 
Temperature and saturation of dissolved oxygen were significantly higher 
at the exploited bed than at the other two sites (Table 12). Though the 
differences between sites in oxygen and temperature were statistically 
significant, the differences may be insignificant ecologically since 



149 



Table 12. Mean physico-chemical parameters measured at a non- 
mussel area, a mussel sanctuary and an exploited mussel bed in pool 26 
of the Mississippi river. Values of a given parameter with different 
letters varied significantly between sites. Unmarked values did not 
differ significantly. 



Site 
Parameter Non-mussel Mussel sanctuary Exploited bed 



Temperature °F 


75.2° 


75.2° 


76. 6 a 


Dissolved oxygen 
concentration (ppm) 
saturation (%) 


9.0 b k 
106. 00 b 


9.3 D k 
109. 75 b 


11. 3 a 
132. 75 a 


Alkalinity (ppm). 


188.1 


188.1 


188.1 


PH 


8.4 


8.6 


8.6 


Conductivity (umho) 


482.50 


471.25 


473.75 



Current velocity (ft/sec) 1.65 1.32 2.91 

Secchi disc depth (inches) 14. 2 a 11.6° 11.9 ab 



150 



dissolved oxygen was at 100 percent or greater saturation each day at 
all three sites and mean temperature was only 1.4 °F higher at the 
mussel sanctuary site. 

Substrate at the non-mussel site was sand; at the other two sites 
the substrate was rock or firm mud colonized with mussels. Substrates 
retrieved with the ponar dredge from the mussel sanctuary and exploited 
mussel bed were heavily colonized with insect larvae tentatively 
identified as caddisflies (Trichoptera) . During the week of sampling, 
adults of these insects were abundant enough along the river banks to be 
a nuisance. 

Fish sampl inq. Fish were sampled using an AC electrofishing boat 
and large and small hoop nets in tandem sets, one tandem set at each 
site. About 125 yards of shoreline were electrofished at each site on 
all days. Electrofishing runs were seven minutes long; half the time was 
spent going downstream, the remaining time was used to go back upstream 
past the same section of shoreline. The nets were set overnight and 
checked each day for three days. 

The species and length of each fish were recorded for each site 
each day. Fish were released at the site. Mean fish length and mean 
number of fish captured were compared between sites. Electrofishing and 
hoop net catches were pooled for statistical analysis. 

Results. The mean daily number of fish caught did not differ 
significantly between sites nor was there a significant difference in 

151 



the mean number of species between the sites. The number of fish species 
caught over the four-day sampling period was highest over the exploited 
bed (N=16), lowest over the sanctuary (N=ll), and moderate over the 
non-mussel area (N=14). Table 13 shows the species captured at each of 
the three sites. We have roughly categorized the species into three 
groups: sport fish, forage fish and "other" species. This last group 
contains both commercial and non-commercial species. 

The mean number of fish of a given species caught daily varied 
significantly among sites in four species: black crappie ( Pomoxi s 
nigromaculatus ) , drum ( Aplodinotus grunniens ) , common carp ( Cyprinus 
carpio ), and emerald shiner ( Notropis atherinoides ) . Black crappie were 
captured only at the non-mussel site. Drum were significantly more 
abundant at the exploited mussel bed site than the non-mussel site. Carp 
were significantly more abundant at the mussel sanctuary site than at 
the exploited mussel bed, but the number of carp captured at the non- 
mussel site did not differ significantly from the number of carp caught 
at either of the other two sites. Significantly more emreald shiners 
were caught in the non-mussel area than in the sanctuary, but the number 
caught at the exploited site did not differ significantly from the other 
two sites. Mean daily catch of the other species did not differ 
significantly between sites (Figure 33). 



152 



Table 13. Species composition at each site in pool 26 of the 
Mississippi River. Presence at a location denoted by 'X'. 









Site 




Group 


Species 


1 


2 


3 




Pomoxis nigromaculatus 


X 








Lepomis macrochirus 


X 


X 


X 




Micropterus salmoides 


X 


X 


X 


Sport 


Morone chrysops 

Morone mississippiensis 


X 


X 


X 
X 




Ictalurus punctatus 


X 


X 


X 




Pylodictus olivarus 


X 


X 


X 




Dorosoma cepedianum 


X 


X 


X 




Dorosoma petenense 


X 


X 


X 


Forage 


Notropis atherinoides 
Notropis blennius 
Hybopsis storeriana 


X 


X 


X 
X 




Alosa chrysochloris 


X 




X 




Aplodinotus grunniens 


X 


X 


X 




Cyprinus carpio 


X 


X 




Other 


Carpiodes carpio 




X 


X 




Lepisosteus platostomus 


X 




X 



153 



■\ 



n m 



2 
o 






o 

Q. 

co 



I 



fc 



S2 



CD 



O 



ID 
Q 



I 
O 
CO 



I 

CO 
UJ 



o 
o 



+_> £ ro It ngu. 

3 r- .- J= .r-,CQ 

t o v) .!Too 

o «_> , « c r^ 



(/> 






I— 5- O. 
.<!>.= _ 

I >>■»-> v^ 



o 

CO 



o 
K 



§ 



8 8 
jeqiunu uee|/\j 



C gj CO or) «» 

■•" ,— H- >■ u. - 

"O V) -i- •*■ .1= 
J) WD w .«U 






Q."D +J 



O) 



V) 



0) 



<« J_ "C ^_ (_) 



o .,=■ -5 <o » ~~ " 
.e i— 4- ■* n 

v> a..,- n i a. 
.i- X C ' f- s- ■ 



**- c 



"O V) 



1/0 



w ™ id w°2 -Sz 

«_ 1- V) 63 S_ 1 00 



3 «fl 



O .« 

w O Q> 



CO 



CO 



111 

CO 
CO 

D 

z 
o 

z 

□ 



CD •*• _ e^ 

°- e ° o +-> a. _c 

v> * « i * i.£=o 

.c M en > o s- 
2-r- i 4-j ' s_ i— 

4- E .c (fl -O O 

O 4_> CO r— ^ Q> 

-.T- S M- 00 J_ 

t i— * _l OS 01 i 

•W M .. ..(OX 

f> </> i2r-o s- So: 
f i a5<-n: a>-ecj 



CD 



a> 



s- o 



oo 



i c <— 

, .O -r- i <+- 

! c+» 4- -o n- 



0) c 



00 



o -E 



*" •<- O. CO ' i-i— 

3 U£ o.^: S- i— 

Dlui+JiUuXl-ra 
H- .r- .p- t_ I/) U E 

u_ s: 2 o i uj x: co 



154 



Gizzard shad were by far the most numerous fish captured (total 
449); the next most numerous species was drum (total 134), followed by 
white bass with a total of 31 fish caught. Total catch numbered sixteen 
or less for each of the other species over the four-day trial (Figure 
34). 

Mean length of drum was significantly greater at the non-mussel 
site than at the exploited bed. Mean length of flathead catfish 
caught over the two mussel bed sites was significantly greater than that 
of flathead catfish caught at the non-mussel site. Mean size of gizzard 
shad caught over the exploited bed was significantly smaller than the 
mean size of gizzard shad caught at the other two areas. No other 
species exhibited significant differences in length between the three 
sites (Figure 35). Over all species, the length-frequency distributions 
were similar for the three sites, though more smaller fish, 
predominantly gizzard shad and drum, were captured over the exploited 
mussel bed (Figure 36). 

Discussion. Over this short trial neither the number of species 
caught per site nor the number of fish caught per site were 
significantly correlated with any of the water quality variables or 
substrate. Over all, there were no significant correlations between 
species length and any of the chemical parameters. There were a few weak 
but significant relationships between the number of fish of a given 
species caught and the chemical parameters; most occurred with species 
which appeared infrequently in the daily catch (Table 14). 



155 




-co a) 

+-> QC r— oo Q. > 

(J 1 U_ ' — ' "I— 



oo 



XI J3 



CD 
l d) 



o-a viT3 t- 

n ™ (8 ' 

N z 



> Q. 
CO 



1? </ »' r " 



rtj 4J CO 



o j_ 
r <-» o 
o .- * « a> £ 
,2 ...*5 = ' « 

T = en 



a> 



1- « rn+J _ ' ^ 



a) 

to 



t «1 -J- 

W r— ry >- 

2 



o 



aj 



<_> CO <_) CXM- 

«j <u :e 3: -r- ifl>*- 
o» oo — ' u_ -C 0) 



.•o 






5ssS|g5 

C T i r- 5 

«« CCQ c^ Cu. 



r— -r- <U <j nj 



• -i- Q. , to 



fc« 



— 2- 



00 <-> 



c v> a> 



-00 -O «• 3 

3 00 O *" (V Si Q. 

^T* #>■» /H /A C .,— * 

- J*: as 

00 O 



3 00 U <" 

a> oo ra oo _ 

•I- •!- I— «J .C 

1LZHJ3 



jequjnu |B}01 



156 




LO 


CD 


N 


N- 


00 


oo 


0) 


r^. 


CO 


CD 


LO 


y— 


1^ 


CO 


(M 


CM 


t- 


T" 


I— 







(seipuj) Mi6ue| ueew 



q 

g 

9 

a. 

s 






co 



GO 



O 

z 



■P E fl fl (OLi. 

=Jr— .-_C .^CQ 

H_ .— -O oo oo ,_, ^ 

•i- W 



tO 






1 (fl OJ O <TJ 



(/) 



O) 



5 s- •- S-Q- 
° (o . a>.c_ 

c a> cu ' i 

O) CO (/j •- 

"O r- l*_ >- u. -O 



«- W-r- 






.' s- 



-= ^ * £ -o £ 

•r- O ^ J= b 00 



r-^ 









to 



£"0 W— I (/, 

.,-5 •* c t3 

00 <fl 



JQ 



oo 



00 



J= ^. CO £ .^ _ 
O O •>- ^ n_ oo 

is c o - ■■ 

ai a} aj 



co O. 



+J CO 



a- .e 
s- a: -o 
(D < ai 
> <_> s- 



3 "O 
I- r— 



-•rZVOI J- 






a: a) i 



<u 



o> ." 



® = „, 

C^i — oo °° (« 

<« c *± n- -a ti_ 

ai ••- c , pf-js 

•r- J- cj o s- q: 

CO n- J*- .« 0) g .» 3 

oo 2Z a> c *" cd o 

a> 00 «3 -r- c c e 

J- •!- a. n» i ■<- t— 
3 wr cc s- r— 

•r- •.- •,- s- oo a> E 

u_ 21 3 o i uj .c oo 



157 



100 
80 
60 
40 
20 


100 
80 
80 
40 
20 





NON-MUSSEL AREA 



BajytliiBanaQEMjnEia 



rir*i =-<± a ' nta I — L. 



0.5 2 3.5 5 6.5 8 9.5 1112.5 14 15.5 17 18.5 20 21.5 23 24.5 28 



MUSSEL SANCTUARY 



nf'IMn.Jffln-raFUmnrL 



rami II HI II lg.1 IHIilM-lill InmHI l.i.nnj,nnm_n I 1. mi i -u ' - I 

2 3.5 5 6.5 8 9.5 1112.514 15.5 17 18.5 20 21.5 23 24.5 26 



0.5 



100 
80 
60 
40 
20 





EXPLOITED MUSSEL BED 



I i-iFltTln-^nNmFlnfil n^. nmn I PI -u - I j. 



0.5 2 3.5 5 6.5 8 9.5 1112.514 15.517 18.5 20 21.5 23 24.5 26 

Length (inches) 



Figure 36. Combined length-frequency distributions for all species at 
the non-mussel, mussel sanctuary and exploited mussel bed sites in pool 
26 of the Mississippi River. 



158 



Table 14. Regressions (r) between total daily number of fish 
captured over the four day period and physico-chemical parameters at the 
three sites. * - Significant regression (P< 0.05). 





Dissolved 




Velocity 


Secchi 


Temperature 


Species 


Oxygen 


PH 


cm/sec 


cm 


Black crappie 


-0.27 


-0.36 


-0.37 


0.48 


-0.33 


Bluegill 


-0.35 


0.10 


-0.25 


0.38 


-0.39 


Largemouth bass 


-0.34 


0.02 


0.32 


0.54 


-0.17 


White bass 


-0.32 + 


0.20 


0.37 


0.55 


-0.17* 


Yellow bass 


0.78 


0.27 


0.28 


-0.33 


0.66 


Channel catfish 


-0.04 


0.18 


-0.18 


-0.45 


-0.42 


Flathead catfish 


0.10 


0.11 


0.53 


0.00 


0.47 


Gizzard shad 


-0.01 


0.09 


0.53 


-0.18 


0.14 


Threadfin shad 


0.09 


0.15. 


-0.26 


0.31 


-0.05 


Emerald shiner 


-0.21 


-0.75 


-0.09 


0.54 


-0.19 


River shiner 


0.00 + 


0.18 


-0.31 


-0.14 


0.13 


Silver chub 


0.74 


0.18 


-0.15 


-0.14 


0.45 


Skipjack herring 


0.47 + 


0.18 


-0.24 


0.06 + 


0.19 


Freshwater drum 


0.58 


0.47 


-0.18 


-0.69 


0.45 


Common carp 


-0.34 


0.26 


-0.39 


-0.32 


-0.40 


River carpsucker 


0.39 


0.21 


0.03 


-0.15 


0.05 


Smallmouth buffalo 


-0.09 + 


0.08 


0.28 


0.23 


-°- 04 * 


Golden redhorse 


0.81 


0.17 


-0.16 


-0.21 


0.66 


Shortnose gar 


-0.29 


0.09 


0.13 


0.01 


-0.19 



159 



With the exception of emeral shiners, drum, carp and black 
crappie, these results do not indicate any striking differences in the 
fish populations between mussel and non-mussel sites. Part of the reason 
for this may be due to the relatively low catch rates of most species. 

Electrofishing took 96 percent of the total catch (614 of 638 
total fish). Thus the apparent lack of difference between sites is not 
too surprising since electrofishing was applied along the shoreline rip- 
rap and boulder habitat which was common to all sites. The hoop nets, 
which could better sample fish associated with benthic habitats, caught 
too few fish for statistical analysis. 

Further field work should attempt to apply a wider variety of 
sampling strategies over a longer period of time. For example a sampling 
period of two to three weeks utilizing longer shocking runs, use of deep 
water shocking methods, and application of a variety of net types such 
as fyke, hoop and trap nets would increase the number of fish captured 
and help increase the number of species captured. Also, since different 
fish species may exhibit temporal variation in their distribution and 
may be attracted to mussel beds during periods of glochidial release, 
samples should be taken during the spring and summer months. 

Identification of sites which are as uniform as possible from the 
standpoint of physico-chemical parameters in conjunction with 
qualification and quantification of the substrates and invertebrate 
forage available at each site will be essential for relating observed 
differences in fish pppulations to the presence or absence of mussel 
beds. 

160